Most people who try to memorize kanji readings hit the same wall. You look up one character and find it has two on'yomi, three kun'yomi, and a couple of exceptions that follow neither. Then you try to hold all of that in your head before you've seen the character used in a single real word. It doesn't stick because there's nothing for it to stick to.

This is one of the most common beginner questions out there. People ask whether they should learn radicals first, then each kanji, then words, or skip ahead and just learn words. The honest answer is that readings were never meant to be memorized as standalone lists. They're patterns you pick up by meeting the kanji inside words you actually use.

Why the list approach fails

Take a common character like 生. On its own it carries something like eight readings. Sei, shou, i, u, o, ki, nama, and more depending on the word. Drilling those eight in a row is close to the hardest possible way to learn them, because a bare reading with no word attached is just a sound with no meaning behind it. Your brain has no reason to keep it.

Now put each reading where it actually lives. 学生 is gakusei, a student. 生まれる is umareru, to be born. 生ビール is nama biiru, draft beer. Same character, three readings, and now each one comes wrapped in a word that means something and shows up in real life. That is something your brain will hold, because it has a hook.

There's research behind why this works. Craik and Tulving showed back in 1975 that people remember information far better when they process it for meaning instead of drilling it as a bare symbol. A reading tied to a word you understand gets processed for meaning. A reading on a list does not.

What worked for me

The more kanji I learned, the more vocabulary came with it. Signs got easier to read, then documents, then words I'd struggled with started clicking because I could finally see the characters sitting inside them. The readings I actually kept were never the ones I drilled from a chart. They were the ones I met inside words I needed. When I was working in Japan and had to read real material every day, the readings sorted themselves out faster than any list ever delivered.

I'm a visual learner, so seeing the character inside the word did more for me than hearing it ever did. That's worth knowing about yourself before you pick a method.

A simpler way to go about it

Learn each kanji through one or two common words rather than its full reading list. Let the high-frequency readings come first, since those are the ones you'll meet again soonest. You'll notice on'yomi tend to show up when kanji combine into compounds, and kun'yomi tend to show up when a kanji stands alone with hiragana attached. You don't have to memorize that rule. You'll feel it after enough words.

Don't try to collect every reading up front. A character that gives you trouble will give it up the third or fourth time you see it in something real.

SimplyKanji is built around this. Every character comes paired with example words, so you see the kanji and then the words it appears in, and the readings arrive carrying meaning instead of sitting on a list to grind. It covers 2,425 kanji and around 6,000 example words organized by JLPT level, and N5 and N4 are free.

References

  1. Craik, F. I. M., & Tulving, E. (1975). Depth of processing and the retention of words in episodic memory. Journal of Experimental Psychology, General, 104(3), 268–294.