Abstract: It is a problem that the fault diagnosis process using wide-area measurement system（WAMS） measurement data is susceptible to bad data.In this paper,the random matrix theory was applied to the information extraction of WAMS measurement data.The fault time determination method based on mean spectral radius（MSR） index and the fault location method based on phase measurement units（PMU） data correlation analysis were proposed,respectively.Through the modeling of the random matrix of PMU data and the mean spectral radius index and the PMU data correlation calculation,the fault time determination based on the MSR index and the fault area location based on the PMU data correlation analysis are realized,which can effectively avoid the interference of bad data.The validity and practicability of the proposed methods are verified by the simulation results of IEEE 39 benchmark system and actual WAMS data.