In this paper, the fault diagnosis and health management of fixed-wing UAVs are investigated based on the deep learning technique. The proposed method includes 5 models: flight data generation model, sample training prediction model based on the Long Short-Term Memory (LSTM) network, prediction model based on the grey model, combined prediction model and health calculation and management model. The real-time output of the health prediction value of the fixed-wing UAVs can be obtained, which makes it possible to take remedial action before the fault occurs. And numerical simulations demonstrate the feasibility of the proposed method.