A couple of questions:
Is the growth due to the log file or the data file in TempDB?
Are you using Snapshot Isolation?
Have you checked for long running transactions?
What have you done so far to diagnosis the issue?
Once you find the issue you will have to look into a potential VLF issue too.