Currently, college students are facing severe employment challenges. To address this, this study uses mental health education assessment data to explore how to optimize the employment guidance and management of college students in order to enhance their employability and psychological adaptability. Decision tree algorithm and improved K-means clustering algorithm are used to analyze the mental health assessment data in depth. Data mining techniques were used to identify students’ employment tendencies and psychological characteristics so as to develop personalized employment guidance strategies. It was found that the accuracy and effectiveness of career guidance can be effectively enhanced through a data-driven approach. Specifically, the accuracy of predicting employment tendency using the decision tree model reaches 85%, while the application of the improved K-means algorithm for clustering analysis of students’ psychological characteristics enables more precise customization of employment guidance strategies. Comprehensive data analysis and psychological assessment can significantly improve the quality of employment guidance for college students. Colleges and universities should pay attention to the use of mental health assessment data, combined with data mining technology, to optimize career guidance and management, in order to help students better adapt to the workplace and enhance their employment competitiveness.