Freestyle skiing aerials is a sport that attracts people's attention and love. The launch speed is very important for freestyle skiing. In addition to its own movement adjustment, it is also affected by environmental resistance factors such as wind speed, wind direction and snow surface friction resistance. In this paper., a freestyle skier's speed quantitative prediction system composed of snow environment measurement system and BP neural network prediction model is designed. The researchers went to the ski resort to make field measurements and used the real-time data measured by the snow environment measurement system and the athlete's own parameters as inputs to the prediction model. Finally., the predicted value of the launch speed is obtained., which is convenient for the coach to provide an effective scientific reference and improve the athlete's competition performance.