In order to grasp the employment psychology of college students more accurately and solve their inner anxiety, the Apripri algorithm of association rules constructs the correlation analysis model of college students' mental health and employment based on data mining. The diagnosis accuracy of association rules for network fault is 98.47%, and the diagnosis time is 0.21s. In the performance comparison experiments of different models, the mean value is above 0.8, the precision is 0.86, the precision is 0.84, the recall is 0.84, and the F1 value is 0.87. It shows that the means of this paper meet the research requirements. In the comparative experiments of different algorithm performance indicators, the accuracy of the mean is 0.87, the precision is 0.85, the recall is 0.84, and the F1 value is 0.88. The means of this paper meet the research requirements.