Abstract: Based on the classic development theory and from the perspective of learning experience， regression and neural network analysis are combined to empirically explore the affecting mechanism of university students learning outcome. Taking account of such factors as individual students and universities， a predication model of student learning outcome is constructed. The findings indicate that student engagement is the most important factor for learning outcome when it comes to individual students， learning resources and campus culture are two major factors for learning outcome at the university level， and learning satisfaction exerts a certain effect on learning outcome in terms of both students and universities. In addition， an integration of academic system and social system could contribute to a better learning outcome for university students.

Key words: student learning outcome, educational data mining, learning analysis, learning experience, student development