We propose an approach to construct an intelligent tutoring system (ITSs) using a knowledge graph (KG) that reproduces the classroom lecture style of human teacher. We propose an approach to divide hierarchical partitioned knowledge graphs into subject graphs, hyper-concept graphs, basic concept graphs, sub-concepts and help-concept graphs. Hierarchical partitioned knowledge graphs can be easily constructed using visualized graphics drawing tools and embedded in the database, the product of the authoring tool. Using the intelligent tutoring system authoring tool, we embed domain model data, basic data of student model, and basic data of pedagogical model into the knowledge graphs. When using the ITS authoring tool, the results are a database and a set of files of various types. The ITS players use the knowledge graph and learning path generator to generate the learning path inherent to the student. The results of the evaluation and analysis of the performance gains of students who studied with the contents of intelligent tutoring using the knowledge graphs were very positive.



