Information and communication technology was currently growing rapidly, one of which was chat or instant messaging applications such as whatsapp, line and telegram.  In october 2020, the majority of instant messaging app users was whatsapp app users, with a total of 2 billion users.  Even though the whatsapp application was in the top ranking and got the highest score, but this could not been used as a measured of satisfaction because there were still negative views on the whatsapp application, some users assumed that whatsapp often had errors when used, then other problems that arise such as the network that the user used was unstable.  To conduct an analysis of this, a sentiment analysis approached was needed to categorize user comments into positive or negative.  This studied used the naïve bayes algorithm with support vector machine in analyzing positive and negative comments on the satisfaction of users of the whatsapp application on the google played store.  From the results of tests carried out on 1500 user commented data, the evaluation of the model used 10 folded crossed validation shows that the leveled of accuracy for whatsapp application user satisfaction based on the naïve bayes algorithm was 70. 40% and support vector machine was 77. 00%, while the auc valued naïve bayes was 0. 585 and support vector machine was 0. 876. From these results, the svm algorithm could been used for researched with the same data characteristics.