This paper mainly discusses the applications of real-time data mining technology in fault prediction of power plant generator. Massive real-time historical data of thermal power plant turbine generator equipment is stored to realize comprehensive quantitative assessment of thermal power plant turbine generator’s online security status and potential failure Early Warning. It is based on the Real-time data mining analysis and modeling techniques.