Aiming at the shortcomings of BP neural network, such as easily falling into local optimum, low diagnostic accuracy and slow convergence speed, a transformer fault diagnosis method based on adaptive differential evolution algorithm and BP neural network is proposed. This method uses differential evolution algorithm to optimize the initial weights and thresholds of BP neural network, and assigns the optimized results to BP neural network for network training. Finally, the optimal network model for transformer fault diagnosis is obtained. The experimental results show that the combined algorithm has higher diagnostic accuracy and faster convergence speed than the traditional BP neural network, and is a more efficient method for transformer fault diagnosis.