Abstract: When the accident occurs in a nuclear power plant, the quick and accurate diagnosis of the accident scenario is of vital importance to take safety measures during emergency response. BP neural network was applied in diagnosis of part of the accident scenarios for AP1000 nuclear power plant. Part of the results of the accident analysis was used as the training data of the network, and the remaining was used as the test data of the trained network to verify its accuracy. The verification result shows that BP neural network is able to diagnose the analyzed accident scenarios with high accuracy.