Abstract: Safety, stability and efficiency, flexible energy flow, and economic and environmental benefits are the basis of a microgrid's low-carbon economic operation, however, with access to multiple types of distributed power sources and flexible loads of multiple interests to the microgrid, the technology of "cloud computing+big data+internet of things+mobile internet+artificial intelligence" is extensively applied, and the network mode and operation of a microgrid are changing. To this end, the key technologies and challenges of AI-enabled microgrid operation are summarized and summarized. Firstly, we introduce the physical architecture of microgrids and summarize the development trend of intelligence, and then we discuss the characteristics of microgrids and the challenges of optimized operation under the goal of a low carbon economy. Secondly, the principles of AI-enabled smart microgrid optimization are explained in terms of decision variables, optimization objectives, constraints, and solution methods. Then, the application effects of AI-enabled microgrid optimization are analyzed and summarized by focusing on typical application scenarios such as renewable energy power forecasting technology, microgrid optimization scheduling technology, carbon trading mechanism and uncertain regulation technology under the deep integration of AI. Finally, the future development direction of AI-enabled microgrid optimization operation is summarized and analyzed to provide a reference for the development of green microgrid technology.