Aiming at the characteristics of RBF neural network which can automatically create a systems dynamic model under the circumstances of unknown parameters

fast track non-linear function and classification capacity.On the basis of the analysis of time signal of breaking coil of high-voltage circuit breaker

a basic method which use RBF neural network to classify faults for high-voltage circuit breaker has been proposed.Compared with the diagnosis result of BP neural network

it shows that RBF neural network has capability of fast convergence

small output error and discrete.By means of the simulation results the effectiveness of adopting this method to diagnose fault for high-voltage circuit breaker has been proved.