The invention discloses a fault reason suspected degree analysis method based on a metering abnormity correlation degree model, and relates to a fault reason suspected degree analysis method. In dailyoperation, many kinds of metering anomalies exist, the situation is complex, and effective analysis, processing and application are difficult to carry out. The method comprises: calculating the association degree of every two abnormal events through classification and screening of historical data, visually seeing whether the occurrence of the two abnormal events has a causal relationship or not,screening results, and providing a basis for subsequent calculation of the fault suspected degree; analyzing and mining a large amount of historical data, and calculating the posterior probability ofa certain abnormality caused by a certain reason; wherein the correlation degree and the posterior probability act together to obtain the fault suspected degree of the concurrent abnormal event causedby a certain reason. According to the technical scheme, fault suspected degree calculation greatly facilitates maintenance, when a power grid system has a fault alarm, on-line operation can be directly carried out, and the reason causing abnormity can be accurately and rapidly obtained by applying the technology.