Abstract: In order to cope with the uncertainty in the large-scale urban power grid system, improve the operability of emergency management theories and methods, and improve the real-time and intelligence of emergency decision-making in the power grid system, this paper proposes a dynamic decision-making method to describe five major functions, including data acquisition, data storage and management, data analysis, decision execution and evaluation, and simulation technology. Firstly, the characteristics of big data in the power grid system are analyzed, the principle of data-driven emergency response decision-making is discussed, and the system architecture and functions of emergency response decision-making methods are deployed from the two dimensions of technical scenarios and the evolution process of emergencies. Then, the research requirements and implementation strategies of three key technologies are expounded, including multi-source data fusion technology, evolutionary modeling technology for emergencies, and intelligent optimization decision-making technology for complex large groups. This paper can provide practical guidance for relevant power companies to establish emergency response decision-making methods for emergencies in the power big data environment.