In the hose-drogue aerial refueling, because the hose is a flexible body, there is a floating phenomenon under the disturbance of complex wind fields such as tanker wake, gust and atmospheric, which seriously affects the success rate of docking. In view of the high accuracy and real-time requirements in the process of modeling the wake flow field of tanker and studying the influence of wake wind field disturbance, the method of combining neural network and CFD is adopted. Based on high-precision CFD method to generate flow field data, BP neural network is used to train the mapping between wind field coordinate parameters and velocity parameters, so as to establish the tanker wake wind field model. Aiming at the problems of poor anti-interference ability, strong dependence on model and difficulty in engineering realization, fuzzy PD control method is adopted to realize the swing suppression of wing-rudder active stabilization drogue under complex wind field interference. The simulation results show that the average floating amplitude in Y and Z directions is reduced by 86.49% and 79.04% respectively, and the flight test results show that the average speed intensity in Y and Z directions is reduced 72.98% and 65.29% respectively, which effectively restrains the floating amplitude of drogue.