Along with the development of the movie industry and the internet, watching movies has become accessible to the public. Nowadays, it is easy to watch movies, and people need to know whether a film is good or not through a collection of movie reviews. The extensive collection of movie reviews spread across many sites makes it difficult for the public to find valid reviews of existing films. From these problems, the public can search for valid and valuable information about film reviews. Therefore, it is necessary to use sentiment analysis on movie reviews to make it easier for the public to find valid and valuable information. The method used in this research is Naive Bayes as classifier and Word2Vec as feature extraction. Word2Vec is chosen because this method can group words with the same meaning in a vector form. Naive Bayes is chosen as a classifier because the method works quickly and is easy to implement. The best model from this research produces an accuracy value of 72.23%.