This paper proposes an algorithm framework for generating pairs of combined test case sets based on particle swarm optimization. When generating test cases, the framework uses particle swarm optimization to try to generate test cases with strong combined coverage capability, and studies the search space. The network environment is described. Undergraduate mental health education system. The system has a clear process and strong practicability, which can reduce the burden of teachers, improve work efficiency, and provide effective support for college students' mental health education. Various analysis methods are used to extract eigenvalues from signal analysis and modeling. The evaluation system has been completed, and the psychological stress evaluation system adapted to the clinical needs of the existing hospital has been completed.