This research is to clarify how to test the thesis topic similarity, make it easy to check the topic thesis, whether it has been made by a student before. In this regard, an important issue that can be raised is how to make a thesis topic similarity test the manual way to be automated. The purpose of this study is a research method similarity of thesis topics using the TF-IDF method. In this research the system has two stages of process, the first mining the text that is categorizing the thesis that has been categorized using the TF-IDF algorithm, which is to read the appearance of each word in the contents of the document. The second stage results from the TF-IDF algorithm are reprocessed with the VSM algorithm. The end result of this program will get the names of documents that have a degree of similarity with keywords.
