In recent years, college student suicide has become the second cause of death on campus after accidents. This project introduces machine learning methods to the mental health monitoring of college students and constructs a model for analyzing the trend of college students’ suicidal tendencies using the SVM algorithm. First, we researched the mental health problems of college students in several colleges and universities and constructed the SVM model by using screened mental health monitoring variables as input features. Its performance is analyzed through model evaluation and comparison, and it is used in the intervention of college students’ mental health problems. Experimental analysis reveals that the SVM model constructed in this paper presents a better performance in analyzing the trend of suicidal tendencies, with a prediction accuracy of 0.935 for different suicidal tendencies, and the results of different assessment indexes are better than the comparison model. After using the model to assist universities in providing mental health interventions for students, the suicidal ideation among sample students decreased by 42.41%. The use of the SVM model to predict college students’ suicidal ideation is effective, and it can be used as a convenient and efficient tool for screening and assessing the risk of suicide attempts to improve the efficiency of suicide prevention.