Guide
Wubi is a Chinese input method that uses character shape instead of pronunciation. For example, in Wubi 86, 林 breaks into 木 + 木: both roots use S, and a final identification key Y completes the code SSY. The letters describe the character's structure; they do not spell its Mandarin reading, lín.
This tutorial starts with the keyboard zones, then works through six characters and a two-character word. You can try the codes in a blank document with your system Wubi input method enabled. Uppercase letters make the codes easier to read here; type lowercase letters without holding Shift.
Wubi requires more study than Pinyin. It is worth considering if you already recognize Chinese characters and want to type from their shapes, especially when you cannot recall a pronunciation. These examples use Simplified Chinese.
Every code and root mapping below uses Wubi 86. Wubi 98 and Wubi New-Century have different root layouts and some different codes. Use a chart and dictionary that match your input method's version; changing versions halfway through learning can make a correct code appear wrong.
The full codes and short codes in this article were checked against the Rime Wubi 86 dictionary. That verifies the dictionary entries, rather than the candidate order or behavior of every Wubi keyboard.
For the full device setup, see Windows Chinese typing or Mac Chinese typing. Once Wubi is selected, keep a blank document open for the examples below.
Wubi roots, called 字根, are the components used to build a character's code. A key can represent several roots. For example, W represents 人 and its side form 亻; S represents 木. Roots are not simply the dictionary radical of each character.
The five zones broadly group roots by their first stroke. In the table, each key is paired with its representative root. Read the keys in Wubi zone order, which differs from left-to-right QWERTY order.
| Zone | Stroke family | Keys and representative roots |
|---|---|---|
| 1 | Horizontal 一 | G 王 · F 土 · D 大 · S 木 · A 工 |
| 2 | Vertical 丨 | H 目 · J 日 · K 口 · L 田 · M 山 |
| 3 | Left-falling 丿 | T 禾 · R 白 · E 月 · W 人 · Q 金 |
| 4 | Dot / right-falling 丶 | Y 言 · U 立 · I 水 · O 火 · P 之 |
| 5 | Turning / hooked 乙 | N 已 · B 子 · V 女 · C 又 · X 纟 |
This is a starting map, not the complete set of roots on each key. Other useful forms include 讠 on Y, 氵 on I, 艹 on A, and 心 on N. Apple's Wubi guide illustrates the keyboard mapping; the Science Press textbook preview shows the Wubi 86 root layout on printed pages 20–22.
Start by locating S 木, J 日, E 月, W 人, V 女, B 子, and K 口. These are enough to recognize the roots in the first five examples.
For these characters, identify the roots in order, map each root to its key, then add an identification key when needed. That final key describes the last stroke and the character's arrangement; it is not another root. The next section explains how to choose it.
| Character | Roots in coding order | Root keys | Final identification key | Full Wubi 86 code |
|---|---|---|---|---|
| 林 | 木 + 木 | S + S | Y | SSY |
| 明 | 日 + 月 | J + E | G | JEG |
| 休 | 亻 + 木 | W + S | Y | WSY |
| 好 | 女 + 子 | V + B | G | VBG |
| 杏 | 木 + 口 | S + K | F | SKF |
| 模 | 木 + 艹 + 日 + 大 | S + A + J + D | None: already four root keys | SAJD |
Try 明 first. Find 日 on J and 月 on E. Its roots sit side by side, and its final stroke is horizontal, so add G: JE + G → JEG. Select 明 from the candidate window, then check the character actually inserted in your document.
Now type 林 and 休. Both end in 木, whose final stroke is right-falling, and both have a left-right arrangement. Their identification key is Y, but their first root differs: SSY for 林 and WSY for 休.
杏 puts 口 below 木. Its last stroke is horizontal, like 明, but its top-bottom arrangement changes the identification key to F: SKF. In 模, 木、艹、日、大 already supply four keys, so SAJD needs no extra identification key.
These examples show why counting visible parts is not enough: you need the Wubi roots and coding rules. Enclosing characters and less familiar shapes also need care with root order and Wubi's last-stroke conventions. Look up their full codes in a matching Wubi 86 dictionary rather than guessing from a Pinyin reading.
In the two-root examples above, the root keys alone do not complete the full code. The 末笔字型识别码, or final identification key, combines two facts: the character's last stroke category and its structure. Left-right, top-bottom, and other arrangements select the first, second, and third key in the relevant stroke zone.
| Last stroke category | Left-right | Top-bottom | Other / mixed |
|---|---|---|---|
| Horizontal 一 | G | F | D |
| Vertical 丨 | H | J | K |
| Left-falling 丿 | T | R | E |
| Dot / right-falling 丶 | Y | U | I |
| Turning / hooked 乙 | N | B | V |
For 好, the last stroke of 子 is horizontal and the character has a left-right structure. That gives G, so 女 V + 子 B + identification G becomes VBG.
Use this table with Wubi's stroke conventions, rather than assuming the last pen movement always gives the answer. It applies to ordinary composite characters that need an identification key. Representative key-name characters use a separate rule: 木 is SSSS, pressing its key four times. S by itself is the root key for 木 inside 林 or 杏; it is not 木's full character code.
You may see a character before you finish its full code. The dictionary can contain a short code as well as a full code: 明 has JE and JEG; 林 has SS and SSY in the checked Rime table. A short code is a shortcut, so it does not necessarily show every part of the full encoding. Candidate choices still need checking.
For a stored two-character word, Wubi 86 takes the first two keys of each character's full code. For example:
KHK → take KH.LGYI → take LG.KH + LG → KHLG.The checked Rime dictionary contains 中国 under KHLG. Word availability depends on the input method's dictionary; if the word is missing, enter its characters separately. Longer words have other coding rules, so do not extend this two-character formula to every phrase.
KHLG, then enter 中 and 国 separately to compare word input with character input.If a code produces several candidates, select the intended character rather than memorizing its position. Microsoft Wubi uses Space to select the focused candidate or a number key to select the corresponding candidate; Enter dismisses the candidate window rather than committing it. Other input methods can behave differently, so use their documented selection controls.
When you get stuck, record whether you forgot a root key, split the character incorrectly, or chose the wrong final identification key. If every example stays as Latin letters, check that Wubi is selected and Chinese mode is active before relearning the codes.
Then try a short Simplified common-character practice. TypeChinese checks the Chinese text you enter; it does not switch your system input method or display Wubi-code hints. Keep your Wubi reference open for unfamiliar characters. Follow the 4-week Wubi typing practice plan for daily root, code, word, and passage checkpoints.
| Question | Pinyin | Wubi |
|---|---|---|
| What do you enter? | Pronunciation | Shape codes |
| What must you learn first? | Pinyin spelling and candidate selection | Roots, decomposition, and coding rules |
| Can you type a character without knowing its reading? | You need a pronunciation or another lookup method | Yes, if you can derive its code and the dictionary includes it |
| What can slow you down? | Choosing among matching characters or phrases | Recalling roots, splitting characters, or looking up codes |
If you already type ordinary messages comfortably with Pinyin, these examples do not create a reason to switch. Wubi can suit sustained Chinese typing or a preference for shape input, but shorter codes alone do not guarantee a higher typing speed. Improving Pinyin input technique remains a useful option.
Use the Chinese keyboard guide to compare other input methods. If Wubi fits your needs, build reliable character input first, then move into words and passages.