In this paper, a smart campus analysis model based on multi-source data fusion technology is constructed with the study of college students’ behavior as the starting point. The K-means algorithm is utilized to mine the campus data of real college students from multiple sources, and through the fusion of multi-source data, a real, intelligent, multi-level student portrait is portrayed in the three dimensions of consumption, academics, and socialization, so as to provide a rational decision-making basis for the comprehensive development of the ideological and political education and student management of the university with the “three-whole-education” approach. The results show that multi-source data fusion technology can significantly reduce the storage space required for data. University X has a low borrowing interest for Civics and Political Science books. Sports are loved by 70% of students, but half of them don’t exercise regularly. Outside spending is a habit among some students, and 20% of them may exhibit abnormal behavior. These individual behavioral characteristics will have a significant impact on the performance of Civics.