The aim of this study is to investigate the attitude of students' pc game and theme preferences (action, strategy, sports etc.). The data were collected from 722 students at the age of 11-14 in Istanbul and was analyzed using by Weka 3.7.0 that is a popular suite of machine learning software. At this point some classifiers and learning algorithms were used for determining which students play which types of games. the study reveals that there is a relation between students PC game types preferences and their demographic characteristics and the independent variables. The software predicts the game types truly in acceptable ratio.