Abstract: Considering that an accurate rescue trajectory is hard to obtain in low altitude and crosswind conditions, the data fusion technique was employed to predict the wind condition at low altitude and, accordingly, to amend the low-trajectory rescue trajectory plan for the aircraft. Taking the international exchange stations within the flying area as the observation points, the numerical weather prediction and interpretation technology based on the unscented Kalman filtering (UKF) was used to build a low-level wind forecasting model by combing the record data sets of wind velocities and directions from the observation points and prediction data sets. Then, the model was used to correct the system error in the original prediction data and produce the modified prediction values about the wind. Finally, according to the principle of velocity triangle, performance parameters such as the rate of climb, cruise speed of the aircraft were combined to estimate the passing time of aircraft at each waypoint. Simulation shows that compared with the results obtained by the traditional Kalman filtering, the root mean square errors of wind speed and wind direction by UKF are decreased by 12.88% and 12.88%;and the initially planned trajectory can be modified more accurately.