Abstract: By using ECMWF high-resolution forecast data, the factors that affect the occurrence of thunderstorm wind are selected from the aspects of dynamics, thermodynamics, water vapor, and unstable energy, the array of multi-dependent variable was constructed, and the ingredient coefficient and its threshold value of the forecasting factor were determined by principal component analysis (PCA). On the basis, the probability forecast product of thunderstorm wind in Sichuan was developed. In 2018, forecasting products of thunderstorm wind were indicative of areas where the probability of occurrence exceeds 65%, and were superior to wind of the ECMWF numerical forecast at the 100 m height. In the cases of test, the effective hit rate of forecast was more than 25%.