The existing abnormal electrical behavior analysis is based on single measuring abnormal event. It has problems such as heavy workload and low accuracy. In order to improveStheSutilizationSofSabnormalSdata and the accuracy of abnormity diagnosis, this paper proposed a method to analyze the fault causes based on the combination of abnormal events. First, a correlation degree calculation method is determined based on the intervals of the date of occurrence. Next, thresholds have been set to identify concurrent abnormal events. Finally, we combined the correlation degree with the probability of the fault causes, and a method to analyze the fault causes has been proposed. The results indicated that this method can improve the accuracy of fault diagnosis.