Plagiarism is the act of recognizing someone else's work as a result of personal work without the original owner's permission work. Plagiarism in the form of documents has happened a lot in this digital era. In response to this, through this scientific work a system will be developed that can be used for detect plagiarism between text documents, namely Rejecting Algorithms with Synonym Recognition. Winnowing Algorithm is a document fingerprint method that is used to detect similarities between text documents using hashing techniques. This algorithm was chosen because Reject is one of the best algorithms to get the value of similarity between document text both in terms of accuracy and the performance. From the results of testing in table 8, the conclusion is that the greater the number of grams and window=2 are used, the process time varies and the percentage of similarity varies, tends to decrease. And also by adding the text processing process and the synonym recognition varies greatly increasing the processing time and decreasing the percentage of similarity.