Abstract: By comparing the shallow layer wind data and the forecasts of Wenchang Space Launch Site on the 26 days affected by typhoons in 2015-2017, the 24 h shallow layer wind forecast capacities during typhoon affecting processes were analyzed.Results show that the average forecast deviation was large. In order to improve the forecast technique, the study firstly analyzed the relationship between typhoon characteristics and shallow wind forecast deviations by using chi-square test method, which identified three key factors: typhoon intensity, quadrant and the distance between typhoon and the launch site. The regression equation of shallow wind prediction for 24 h at launch site was then established using real time data from 2015 to 2016. It was found out that the distance between typhoon and the launch site and the numerical prediction data are the main factors of the regression forecast equation, while the contribution of typhoon intensity and quadrant can be ignored. The verified result of the regression equation by using the data of 2017 shows that the forecast accuracy rate could be improved by reducing the standard deviation of the forecast deviation. This study could provide a quantitative reference for the shallow wind forecast when typhoon affects the site.