The fuzzy computer model of the teacher-knowledge-students system is proposed and its behavior under various conditions is investigated. The factors complicating the study of didactic systems is identified, the cognitive network is built that models the learning process, and its main characteristics is listed: the complexity of educational information, student's motivation, his level of understanding, class attendance, learning activity, home assignment etc. By specifying triangular membership functions, fuzzification of the input values is carried out. The computer program is written that uses a logistic function to calculate the student's average grade and the probability of receiving corresponding mark. To determine the parameters of the model, it is adjusted, during which coefficients are selected. The simulation of student learning with various input parameters is carried out. Also another model is considered, which takes into account the influence of student motivation on the amount of homework and activity in the classroom, and the results of its work are analyzed.