China Itinerary: Choose the Shape Before the Cities

The four shapes

Single hub

One city, all nights in one hotel, day trips radiating outwards. The cheapest in transit and the highest in sightseeing per day. It wins when the trip is short, when the party includes children or older travellers, and when the city itself is dense enough to fill the time.

Two-hub line

Three nights, one move, three nights. One transfer day, two bases, and a genuine change of scenery. It wins when the two places are different in kind rather than degree. It loses when the second city is a smaller version of the first, because you have paid a day for variety you did not get.

Open-jaw corridor

Arrive in one city, leave from another, travel one direction through the middle. This is the shape that adds distance without adding a wasted return leg. Its constraint is the international ticket — if the airline will not sell you an open-jaw itinerary at a sane price, the corridor disappears and you are back to a loop.

Regional loop

A circuit that returns to where it started, usually by rail with a few road segments. This is the right shape for longer trips in the south-west and the south, where the interesting places are scattered around a region rather than lined up along a corridor. Its weakness is that every loop ends with a return leg that earns nothing.

The arrival city decides half the trip

Which airport you land at is not a detail, it is the first structural decision — and it is usually made for the wrong reason, by whoever had the cheapest fare.

The arrival city sets your first region, your first base, and the direction of everything that follows. Land in Beijing and the natural continuation is west along the historical spine or south along the coast. Land in Shanghai and the natural continuation is the eastern cities and the water towns. Land in Chengdu and you are in the south-west, where the distances behave differently and rail stops doing all the work.

It also sets your first day. A city with a fast, simple airport-to-centre connection gives you an evening on arrival; a city where the airport is far out and the traffic is heavy gives you nothing. If your flight lands in the evening, the difference between those two cities is half a day.

So reverse the usual order: look at arrival and departure options first, choose the pair, and only then fill in the middle. This is also how you discover whether an open-jaw corridor is available to you at all.

The rail spine, and where it stops helping

China's high-speed network is the reason a multi-city itinerary is possible at all. It is fast, punctual and dense across the eastern half of the country. But it is a spine, not a blanket, and the itinerary has to respect where it ends.

In the east, rail does everything: the major cities are connected in a chain, stations are in or near the centre, and a move can genuinely be a morning. In the centre and the north-west the network thins and the useful connections become fewer, which means the shape of the trip has to follow the lines rather than the map.

West of a certain point, rail stops being the answer and flying takes over — and flying changes the arithmetic, because airports are outside cities and the door-to-door time includes two transfers. A day that contains a flight contains very little else.

Two operational rules follow. First, choose the corridor where the fast trains already run rather than inventing a route and hoping a train exists. Second, book the rail move before you book the things that hang off it; seat inventory on the popular departures is real. Our pages on using China's high-speed rail and on booking train tickets cover the mechanics.

There is also a genuine rail-versus-flying trade-off on longer legs, and it is not decided by speed alone — it is decided by where the station and the airport sit relative to where you sleep. That comparison is set out in rail versus flying.

Sequence rule: never move on the day you arrive somewhere important

The single most useful sequencing rule is this — transfer days belong to places you have finished, never to places you have come to see.

If you move in the morning and arrive at midday, you have an afternoon, and that afternoon is at the mercy of a delayed train, a slow taxi queue and a hotel room that is not ready. If the thing you came for is a timed admission that afternoon, you have built a plan that fails in the most ordinary way possible.

So put the move at the end of a stay, let the arrival day be an evening, and start the real sightseeing the next morning. That rhythm costs nothing and removes an entire class of failure.

A second rule goes with it: keep the first full day in any new city light. You will be slower than you think — finding breakfast, working out the metro, orienting yourself — and a light first day is what makes the second day good.

Why the famous three get picked, and what they cost

Beijing, Xi'an and Shanghai appear in almost every first itinerary, and the reason is legitimate: they are the shortest route to the country's most recognisable set of monuments and skylines, they are connected by the fastest rail, and they are easy to explain to the people you are travelling with.

What the standard version understates is the cost in days. Three cities in a line means two moves, which means two days that contain a station instead of a sight. In a ten-day trip that is fine. In a seven-day trip it consumes the reserve and the breathing room, and the trip becomes a delivery schedule. That arithmetic is worked through in what one week actually fits.

There is a cheaper version of the same idea: keep two of the three as bases and treat the third as a day trip or a single overnight. You still see it, you still tell the same story at home, and you keep a day.

Adding a region: the cost is re-entry, not distance

When people add a fourth or fifth stop, they usually estimate the cost by distance. The real cost is the re-entry — every new region means a new base, a new set of local logistics and a transfer day, whatever the mileage.

That is why a nearby second city is often more expensive than a distant one on the same rail line: the near city requires a hotel change, a re-orientation and a packing cycle, while a stop on the way to somewhere else costs only the stop.

The practical test is to count bases, not cities. Two bases in ten days is comfortable. Three is workable. Four is a trip about hotels. If a proposed stop does not add a base — because it sits on the line between two places you were already going — it is nearly free, and those are the stops worth taking.

Knots and dead ends

A few structural traps recur often enough to name.

• The return-to-origin trap. A loop that ends where it began often has a final leg that is pure cost. Unless the return leg is cheap and short, an open-jaw corridor is better.
• The one-way problem. Self-drive is restricted for foreign visitors in most of the country, and one-way rentals are not a general solution, so road-based loops have to come back to their starting point. Plan the loop knowingly.
• The hub-and-spoke illusion. A city that looks central on a map is not necessarily a good base if the day trips from it each take a long morning. Check the actual first and last mile, not the straight-line distance.
• The double-back. Two stops that are far apart with a third between them, visited in the wrong order, add a whole leg for nothing. Sort by geography before you sort by interest.
• The border day. A day trip that crosses into a neighbouring region has its own document rules. Confirm eligibility before you build a day around it.

Build it in this order

1. Fix the length. Not the ideal length — the booked length, based on your flights and your leave.
2. Choose the shape from the four above, using trip length and the number of bases you are willing to carry.
3. Choose the arrival and departure cities together, and check what tickets are actually sold for that pair.
4. Place the base or bases so that no base requires more than one transfer day.
5. Sort the stops geographically, then reorder within the geography by interest.
6. Drop one thing. Every good itinerary has one deliberate omission in it, and choosing it in advance is what keeps the rest unhurried.
7. Add the reserve day, unbooked, in the middle rather than at the end.
8. Book in dependency order: entry permission, flights, rail, timed admissions, hotels, local transfers. Our pre-arrival checklist walks through what each of those involves.

What the route diaries leave out

Personal itineraries are useful for two things: they show you what a place feels like at a given pace, and they give you an honest sense of how tired you will be. They are unreliable for structure, for three reasons.

First, survivorship: the routes that get written up are the ones that worked, and the ones that lost a day to a missed connection rarely get published. Second, the writer's constraints are invisible — a route that fits a solo traveller with hand luggage does not fit a family with two suitcases and a stroller. Third, the costs quoted are the writer's costs in the writer's season, which tells you very little about yours.

Use them for texture and pace. Use structure from first principles: shape, bases, moves, reserve. Our China travel guide starts from trip length for exactly this reason, and choosing where to stay deals with the base-level decisions that a route map hides.

A good itinerary is not the one with the most places on it. It is the one where the number of moves matches the number of days you were willing to spend moving — and where everything else was chosen after that, not before.