In order to solve the problem of distribution fault diagnosis in case of misreporting or failed-report of fault indicator information, the characteristics of the fault indicator are analyzed, and the concept of the minimum fault judgment area of the distribution network is developed. Based on which, the mathematical model of fault indicator fault diagnosis is evaluated. The characteristics of fault indicator signals are analyzed. Based on two-in-three principle, a probabilistic fault indicator combination signal processing method is proposed. Based on the combination of the minimum fault judgment area model, the fault indicator combination signal and the interdependence between the fault indicators, a fault diagnosis method based on maximum probability is proposed. The method is based on the similarity between the simulated fault signal and the real fault signal, and the detailed formula is given. The method has good fault-tolerance in the case of misreporting or failed-report of fault indicator information, which can more accurately determine the fault area. The probability of each region is given, and fault alternatives are provided. The proposed approach is feasible and valuable for the dispatching and maintenance personnel to deal with the fault.