These days, e-learning systems have become widespread.Many educational organizations such as universities and enterprises use e-learning systems for education of students and new employees. In universities, some lectures use e-learning systems. In that case, it is often the case that some students do not concentrate properly on learning, and teachers may not be able to grasp situations whether students intensively studied or not. As one of solutions to the problem and to help teachers grasp students' behavior, we constructed an e-learning system that judges whether or not students study appropriately with concentration using Kinect and warns them to concentrate when they do not. In the system, "Face Tracking" (face recognition function), one of the functions of Kinect, is employed for the e-learning system to check whether the learner's face and eye direction are looking forward and whether the eyes are open. In this research, experiments are conducted to improve the accuracy of detection by changing the threshold to determine whether a student is concentrated or not. Additionally, whether the mouse does not move for a predetermined time or more is checked. In order to show the usefulness of the system through experiments, the accuracy of detecting concentration is examined, and the growth rate of the grading points with warning is compared with that without warning. Furthermore, to see the feeling of students, the questionnaire about the system after the experiments is conducted.