This paper combines association rules and collaborative filtering algorithms to build a course recommendation method that considers personalized learning characteristics on the innovative development platform of Civic and Political Education in colleges and universities. The related algorithms in the association rules are used to visualize and analyze the association of the 10 main contents of the Civic and Political courses and give the combination of the contents of the Civic and Political courses. By analyzing the mastery and interest of students on the platform, we push the combination of courses that meet the personalized needs of students according to their knowledge, ability and interest. The constructed personalized teaching platform is used in actual teaching, and the application of association rules in course analysis and personalized push function is analyzed to prove the effectiveness of the platform. The impact of the platform on teaching is evaluated by comparing the students’ performance under platform learning to that under traditional teaching. The results show that the average scores of the Civics course content of group D1 are all above 80, and the significance value of the difference between group D1 and D2 is 4.21% < less than 5%, indicating that there is an obvious achievement difference between the two groups. In the recommended course content of Student 2, the score of the combination course A4 and A7 was 8.2158, and the learning interest rating was 17.326, which was 0.769 higher than the sum of the interest ratings of the two courses alone.