I don't use kate, but I don't think it's an encoding problem. How did you read the file and get garbage?
You can use the "file" command to check a file's encoding.
This would return
filename: UTF-8 Unicode text
On Japanese(chinese and korean too) languages properly encoded in unicode.
If it's not correctly encoded maybe you have an input problem... Try another editor to check?
If it's correctly encoded, then the problem should be the method in which you are reading the file, I guess.
If you checked the text in a browser remember to put the
<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="zh" lang="zh">
<meta http-equiv="Content-Type" content="text/html;charset=UTF-8" />
Telling the browser the language and encoding information to view the page properly.
I've only written pages in Japanese but they should be kinda the same method.