The application of the BP and RBF artificial neural networks in energy characteristics prediction of centrifugal pumps was summarized. The structure and characteristics of the two artificial neural networks were introduced in detail. The models of BP and RBF artificial neural network were established respectively to predict the centrifugal pump energy characteristics. The characteristics data of 57 centrifugal pumps were used to train the two models, and the data of the other 6 centrifugal pumps were used to test the two models. The study shows that the prediction results of the two networks are closer and the trends of prediction results are the same for the two networks. The precision of BP network is a little higher than that of RBF network. The head average prediction discrepancy for BP network is 3.85% and the efficiency average discrepancy is 1.39% points. The head average prediction discrepancy for RBF network is 4.79% and the efficiency average discrepancy is 3.43% points. The prediction time of RBF network is only half the time of BP network.