That is not surprising, you're using different code pages when editing/compiling and running the program, so your Chinese characters come out as rubbish.
If you insist on writing Chinese in a non-Unicode program (which is somewhat unwise), you need to tell the compiler that you want to do such a thing, the switch for that would be something like -finput-charset=window936 for gcc. If you don't tell the compiler anything, it assumes UTF-8 encoding.
Otherwise, leave everything as it is in Code::Blocks and write a proper Unicode program (including L"" strings and the correct #defines for your toolkit/os) and you will have no such problems.