Abstract: As an important research and application area of artificial intelligence （AI）, intelligent tutoring system （ITS） is the natural direction for human digital education in 21st century. Although it has been studied deeply for over 30 years since 1970’s, in some existing intelligent tutoring systems, domain dependent property of knowledge representation and reasoning and/or lack of dynamic connection between system construction and system application and among different modules of the systems are obvious, resulting in poor function of student model, and restricting the intelligence of the system for individualized instruction. Based on knowledge representation approach with SC grammar formalism and with knowledge tree mapping approach, a dynamic, real time, interactive augment knowledge tree model （AKTM） with self adaptive ability is put forward innovatively, realizing both in content and function almost all the classical four modules of ITS seamlessly and dynamically. The model consists of SC grammar based knowledge representation, knowledge tree architecture, knowledge tree mapping, knowledge concept tutoring cycling, etc. With dynamic construction of the knowledge tree and the mapping of the knowledge concept structure and its semantics, board semantic relationship of the complicated knowledge of the ITS is realized by structural and hierarchical attributes annotation and hierarchical indexing of the knowledge concept, enabling knowledge processing, calling, storage and maintenance to be carried out dynamically and systematically in the meanwhile. For instantiation design and application of the model, an applicable multimedia ITS system is designed with intelligent properties such as dynamic tutoring and learning, man machine interaction, domain free, self adaptive, etc.