This study focuses on the mechanisms through which educational building spatial experiences influence students’ mental health, employing empirical research methods for a systematic analysis. A sample of 471 valid responses from students at a certain university was selected. Data collection was conducted using a self-designed spatial perception questionnaire, a well-validated educational building space evaluation scale (comprising five dimensions and 31 items: sense of security and comfort, sense of belonging and identity, sense of control and autonomy, restorative and stress-reducing effects, and social support and connection), and the standardized Student’s Clinical Lottery Scale (SCL-90) for mental health. To address the issue of potential missing response variables, the study employed full information multiple imputation (FIMI) and expectation-recurrent least squares (ERLS) for robust handling. Students’ overall experience of educational building spaces was positive, with all item mean scores exceeding 3.5. Factor analysis extracted five common factors, explaining 89.231% of the cumulative variance, thereby validating and refining the multidimensional structure of spatial experience. Correlation analysis revealed significant negative correlations between the spatial experience factors and mental health issues, with r ranging from -0.228 to -0.082. Multiple regression analysis (R² = 62.3%, p<0.001) further quantified the effects, with a sense of security and comfort (β=-0.304), social support and connectedness (β=-0.261), and a sense of belonging and identity (β=-0.281) being the strongest predictors of mental health status. This indicates that enhancing the sense of security, social support, and belonging in educational building spaces can effectively reduce the risk of mental health issues among students.