Abstract: In order to improve the fault early warning ability of substation, the fault early warning system based on big data analysis is studied. In this paper, the overview, fault problems and challenges of a medium-sized 220kV substation are studied, as well as the application of big data analysis in fault early warning. The design of the fault early warning system based on big data analysis was analyzed, including the system architecture design, fault early warning model construction, system functions and characteristics. Finally, the system implementation process and operation effect evaluation are discussed, and technical feedback and improvement measures are given, which effectively improves the accuracy of substation fault warning, reduces the false alarm rate and false positive rate, comprehensively covers and timely discovers the potential failure risk of equipment, strives for processing time for operation and maintenance, and ensures the stable operation of substation.