The e-learning of the Faculty of Engineering, Muhammadiyah University of Yogyakarta has been so intensively used in the teaching and learning process. One of the benchmarks used in determining a student’s final grade is to take the values that exist in e-learning. This study aims to predict the student’s final grades by utilizing the data mining process and using the naive bayes algorithm. The results of this study provide information that can be taken by students and lecturers to make improvement in the teaching and learning process to improve student final grades and maintain a satisfactory student final grade until the lecture is complete. The research method used begins with the study of literature, data collection, data selection, data cleaning, data transformation and implementation with rapidminer and if successful immediately make conclusion that produce information. The results of the calculation of the prediction of the final grade of students there are 1 courses that get a lot of unsatisfactory grades with an accuracy rate of 93.75%, the higher the accuracy value, the predicted final value is close to the actual value