Purpose: The purpose of this study is to build a personalized learning intervention system, which can support students' personalized learning, improve teachers' teaching efficiency and students' learning effect. Design/methodology/approach: The research proposes a personalized learning intervention method based on a collaborative filtering algorithm and knowledge map. The application of knowledge map makes learning content organized, and the use of collaborative filtering algorithm makes it possible to provide personalized learning recommendations for students. This personalized learning intervention system can monitor students' learning development and achieve the combination of personalized and efficiency. For the study, 152 seventh graders were assigned to a control group and an experimental group. Traditional learning intervention was used in the control group, and individualized learning intervention was used in the experimental group. Findings: SPSS was used for data organization and analysis. The effectiveness of the personalized learning intervention system is verified by quasi-experimental research, and the influence of the system on students' learning effect is discussed. The result found that personalized learning interventions were more effective than traditional approaches in improving students' achievement. However, for students of different learning levels, personalized learning intervention system has different effects on learning confidence and learning anxiety. Originality/value: The personalized learning intervention system based on the collaborative filtering algorithm and knowledge map is effective in improving learning effect. And, it also has a certain influence on students' psychology.