Abstract: The power system is the key to the operation of national economy and has gradually evolved into a cyber-physical-social system. The power system has many difficult characteristics such as high dimensions, tight coupling, uncertainty and complexity. Intelligent analysis of the power system is a complex multi-objective optimization problem. Curse of dimensionality, poor interpretability and fuzzy state remain some major challenges for analyzing in such extra-large-scale systems. In this paper, the framework of the power system enhanced intelligence analysis（PSEIA） was studied on the basis of previous studies of artificial intelligence. It was proposed that the PSEIA was a human-machine hybrid intelligent operation form by "service-driven". Then, the feasibility of the PSEIA was analyzed and its basic theory was also discussed. Finally, in order to actualize the PSEIA applications in power system, the technical difficulties of the PSEIA were investigated. This study is intended to provide reference for the future development of artificial intelligence technology in the power grid.