The digital level of power station has been improved, and the control card used to maintain the normal operation of the electrical system has been produced. However, there are few studies on the maintenance and diagnosis of the card, it has a waste of cards and requires high-level maintenance. In this paper, this kind of control card is taken as the research object. With the help of the neural network model, a systematic intelligent diagnosis model is formed. On this basis, the intelligent diagnosis system is developed to realize automatic power-on test, automatic data acquisition and transmission. Through the analysis of upper machine processing directly diagnosis, reduce the workload of maintenance personnel, improve the value of test data, and promote the electric system intensive management in the field of equipment maintenance.