The invention relates to a spacecraft attitude control system fault diagnosis method based on a BP neural network. Sensor data are collected when a spacecraft attitude control system model does not break down, the BP neural network is trained according to the data, and a fault diagnosis model is established; real output of a spacecraft attitude control system model sensor is different from the output of a fault diagnosis model, and a residual error is obtained for fault diagnosis; and the residual error is compared with a set threshold, and a spacecraft attitude control system is subjected tofault diagnosis. According to the spacecraft attitude control system fault diagnosis method based on the BP neural network, a neural network method is used for processing nonlinear problems, comparedwith traditional control system state observers, a state observer applying neural network design has high adaptivity, the very important method using the neural network for observing and identifying anonlinear system is achieved, and the high practical application value is achieved.