The content of undergraduate Data Mining and Machine Learning course is abstract and difficult to understand. The traditional teaching materials of colleges and universities are mainly lecture teaching. At present, a large number of case-based and project-based teaching methods have been proposed to reduce the learning difficulty, but they still have the disadvantages of abstraction,
not easy to implement and lack of analysis of algorithm principles. In this paper, the process generation heuristic teaching is proposed for the first time in the Data Mining course, and the prototype of students' life experience is gradually formalized into a rigorous data mining algorithm, which reduces the learning difficulty, contributes to students' deep understanding of the algorithm and achieves better teaching results, on the basis of ensuring the correctness and scientificity of subject knowledge teaching, helps students generate subject literacy and subject ability.