Personal network teaching has become a hot topic and an urgent task as well in the innovation of modern education. Data mining provides a strong support for personalized network education.This paper mainly deals with applying data mining into network education, meanwhile, building a common teaching model. This model can trace the study process of students, to create a study log for each student. Also, this model can modify the teaching strategies according to information of each student. Student will get his or her personal education. During the teaching, the system will generate evaluation of student's score and the ability of cognitive automatically. The system will give learning advices which generated from reasoning system.This paper proposes three points. First, the knowledge is organized by using tree-graph structure. Second, fuzzy reasoning are adopted in modifying teaching strategies. Last, the interests of students are found by using improved Approri algorithm. |