The safety and stability control system (safety control system) has many control nodes and a long control chain. If it refuses to act or malfunction due to a fault, it will cause serious harm to the operation of the power grid. The fault diagnosis of the security control system is the basis for the safe and stable operation of the power system. The fault diagnosis of the existing security control system mainly relies on the technical personnel to assist in identifying the abnormal cause according to the communication message, and it is difficult to diagnose the faults of each link of the security control system in real time. To this end, the existing links and causes of safety control system faults are analyzed, and the characteristic quantities of safety control system faults are extracted; then a safety control system fault diagnosis model based on deep confidence network is established, and a safety control system fault diagnosis method is proposed. The fault diagnosis method is verified by selecting the operation fault samples of the safety control system.