In order to improve the operation safety of high voltage power supply, an intelligent diagnosis method of overload fault of high voltage power supply based on deep confidence network is proposed. Firstly, the mapping relationship between the operation state of high voltage power supply and fault signal is expressed by deep confidence network, and the overload fault characteristics are extracted. Secondly, the linear expression function of feature data in high-dimensional space is constructed, and the fault diagnosis problem is transformed into an objective function under constraints. Finally, the objective function of fault diagnosis is constructed, and the objective function is solved under the constraints of KKT to complete the intelligent diagnosis of overload fault of high voltage power supply. The experimental results show that compared with the traditional diagnosis methods, the diagnosis accuracy and efficiency of this method are significantly improved, and the diagnosis accuracy is always maintained at more than 90%.