China Travel Guide: A Checklist You Can Actually Use

What this guide refuses to do

It will not rank attractions. Every large guide does that, and the ranking is the least transferable part of the advice, because the right answer depends on your month, your entry city and how many nights you actually have. It will not describe scenery you can see in a photograph. And it will not pretend that planning is a matter of collecting enough information: at some point more reading stops helping and starts delaying decisions that have deadlines.

What it does instead is treat a China trip as a set of dependencies. A dependency is a task that becomes impossible, more expensive, or much harder if you do it after something else. There are six chains, they overlap, and almost every expensive first-trip mistake is a dependency that was discovered too late.

Begin with dependencies, not attractions

The instinct is to start with places, because places are the pleasant part. The problem is that places are the output, not the input. Your entry point, your dates and your documents determine which places are even available, and all three of those decisions are cheaper to make first.

Start with four fixed facts and write them down before you look at a single photograph. Which airport do you arrive at, and which do you leave from. What are the exact dates, including the day you land and the day you fly out. Which passport is each traveler holding, and what does that passport currently allow. And what is the total budget for transport inside the country, which is the number that decides whether the trip is one region or three.

Everything else is downstream. A route built before those four are fixed will be rebuilt, often more than once.

The dependency chain

Notice that five of the six chains end in a test rather than a purchase. That is the point of the table. Setting something up is not the same as verifying it works, and the verification is the part people skip.

Use the downloadable master checklist

The printable table has one row per task and four columns: task, deadline relative to departure, done, and verified. The verified column is deliberately separate from done, because the gap between those two is where trips fail. A VPN is done when it is installed and verified when a blocked service loads. A payment app is done when a card is added and verified when a small purchase clears. A hotel is done when it is booked and verified when the Chinese address is saved and screenshotted.

Print it if you are the kind of planner who needs it out of the browser, and keep it open in a tab if you are not. Either way, the last column is the one that earns its keep.

Documents and route

Entry rules are the single most volatile part of planning a China trip. Visa-free arrangements, transit permissions and short-stay schemes expanded and changed repeatedly between 2024 and 2026, and they differ by nationality, by port of entry and sometimes by the region you intend to visit. Two consequences follow. First, check the rule for your specific passport and your specific entry point, not for tourists in general. Second, re-check it close to departure, because a decision made three months out can be out of date by the time you fly.

Build the route only after the entry question is settled. Then apply the two structural rules that save the most time. Keep the number of internal moves low; every move costs at least half a day once stations, transfers and check-in are counted. And put the hard-to-book sites on fixed dates early in the plan, then let the flexible stops fill the gaps around them.

One more document habit: keep a photograph of your passport photo page, your visa or entry confirmation, your insurance policy and your first hotel address in a single offline album. When something goes wrong at a border or a police station, an album of screenshots resolves it faster than an email account you cannot open.

Payments, phones and language

Payment in China is mobile-first. Two apps cover nearly everything, both accept foreign cards in principle, and both can fail at the linking step for reasons that have nothing to do with China: a card issuer that wants a verification call, a phone number that no longer receives texts, a daily limit set years ago on a student account. Test the link before you travel and carry cash as a second layer.

Phones need a plan that works around the block on Google, Gmail, YouTube and the major social platforms, because those are blocked on local networks. Either a travel eSIM or roaming plan that routes traffic outside the country, or a local SIM plus a VPN installed and tested at home. Do not plan to install a VPN after arrival; the download itself may not load.

Language is a solved problem if you set it up properly. Install the offline Chinese pack in your translation app, learn about ten phrases, and rely on showing a screen rather than speaking a sentence. The highest-value single feature is camera translation on printed menus and station signs, and it works without a connection once the pack is downloaded.

Bookings and final checks

Three booking types behave differently and should be handled in different ways. Rail is bookable online in advance and ties the ticket to your passport, so names must match exactly and you should book the seat rather than queue on the day. Major heritage and museum sites increasingly use timed reservations that release on a rolling schedule and can sell out well before your date, so these need a calendar reminder rather than good intentions. Hotels are the flexible layer; book the ones near transport rather than near the sight, because you will use the station twice and the sight once.

The final checks, in the last week before departure, take an hour. Re-verify entry eligibility. Re-verify the two timed reservations. Re-test the payment app with a small transaction. Re-test the VPN or eSIM on a network that is not your home wifi. Download offline maps for each city on the route. Save every hotel address in Chinese. Confirm that your bank knows you are travelling, and that your insurance emergency line is saved offline.

The four things that cannot be fixed after you land

Entry eligibility, because the decision happens at the departure gate. A VPN or roaming route that depends on a download, because the download may not be reachable. A payment card link that depends on a verification call to a number that does not work abroad. And a timed reservation for the one site that justified the trip, because the inventory is gone.

Everything else on this checklist degrades gracefully. A missing translation pack means slow conversations, not a broken trip. A forgotten adapter is a shop away. A hotel in a slightly wrong neighbourhood is a taxi ride. These four are different in kind: they are binary, and they are decided before you board.

That is the whole argument for planning China this way. Not because the country is difficult, but because a small number of tasks are irreversible, and the rest of the planning exists to make sure you have room to fix the reversible ones.

Two worked examples of the chain

The table is abstract until you run it twice, so here are two shapes that follow it.

A ten day first trip, one region. Two cities, one base each, one day trip from each, one recovery day. The chain runs: entry eligibility checked, then flights, then two timed reservations pinned to specific dates, then rail between the cities booked to match those dates, then hotels chosen near stations rather than near sights. No internal flights, no more than two long moves, and the recovery day sits on day seven where energy runs out. The failure mode this shape avoids is the four-city version of the same ten days, which has the same sights and roughly half the actual experience.

A two week trip that includes a high plateau. One plateau stop of two nights, one lowland anchor, one base city for the final three nights. The chain changes at the front: the plateau dates get fixed first, because altitude tolerance and weather both want a spare day and because the transfer to it is the longest single move on the route. Everything low-altitude is booked around those two dates. The failure mode this shape avoids is putting the plateau in the middle, which means arriving acclimatised to nothing and leaving on a flight while still short of breath.

Both shapes share one property: the hard, irreversible bookings happen first, and the flexible stops fill the remaining space. That ordering is what makes a route survive a delay. A route built the other way, from the pleasant decisions outward, has no slack left when the first train slips.

A two week countdown that fits real life

Two weeks out: settle entry eligibility and the two timed reservations, and book rail for the fixed legs. Ten days out: choose the data route, install and test everything that needs a download. One week out: link and test a payment method, save Chinese addresses for every hotel, download offline maps and the translation pack. Three days out: recheck entry rules, recheck reservations, tell the bank, save insurance and embassy contacts offline. One day out: screenshot the first hotel, the first train and the first transfer, and charge a power bank.

That is three hours of work spread across two weeks, and it removes almost every failure mode that costs a day. The attractions take care of themselves once the dependencies are satisfied.