These days, relying on the traditional examination process, especially in high schools, is no longer the best solution for educating students, engaging their study attitudes, and keeping them improve self-study ability. The intelligent tutoring system is designed to aid teachers in multidisciplinary constructions of teaching materials. In this paper, the authors look at a real-world problem by proposing a system that helps high school students do multiplechoice examination better. First, the authors receive a dataset from a high school regarding the informatics course examinations. We then perform several machine learning algorithms to select one deployment candidate based on its high prediction performance. We have achieved a good accuracy of 98:93%0:25. Third, we integrate the learned model to a selfdeveloped intelligent tutoring system.