With the curerent development of technology, online games have become a product of technology that is very well known, especially among children and adults. One of the popular games in Indonesia is Player Unknown’s Battleground Mobile or commonly known as PUBG Mobile. The success achieved by PUBG Mobile would not exist without reviews from its users. Sentiment Analysis is known as a method for analyzing user’s opinion that is suitable for discussing reviews on the PUBG Mobile application. To find the best results, this study uses several classification method, namely Random Forest, Naïve Bayes, Logistic Regression, and SVM which will be compared to find out which classification method has the best performance. The dataset used in this study was taken from user reviews of the PUBG Mobile Application on the App Store and in English language. In this study, it was found that the Logistic Regression classification was more accurate the the Random Forest, SVM and Naïve Bayes classification with the accuracy result and the F1 score of 78%.