Studies of regional labor markets reflect the presence of a shortage of personnel with a set of competencies that meet the modern tasks of enterprises. One of the reasons for the identified shortage of specialists with the required qualifications is the inconsistency of the recruitment and content of educational programs, the imbalance in the graduation of students by higher educational institutions. In order to meet the needs of the regional labor market for highly qualified specialists and predict their graduation, it is necessary to identify the value orientations of the choice of bachelor’s degree programs by applicants, as well as the admission of talented master’s students who have shown a propensity for professional activity, in which they have received basic knowledge. This provision determines the relevance of studying educational trajectories in relation to the characteristics of applicants, as well as economic incentives for their behavior in the form of expected income after receiving higher education. The research approach presented in this article involves combining an agent-oriented model with an econometric model to maximize the expected salary of students. At the first level of elaborating a multi-level model, an econometric model is used, which allows adapting educational trajectories to the interests of students. At the second stage, the agent-oriented model describes the behavior of students in the general sample. The advantages of the presented approach are the possibility of calculating the optimal options and scenarios for students’ training, taking into account the success of graduates in the regional labor market. To determine the probability of a student changing the educational trajectory, a fuzzy logic model of the Mamdani type was developed.