Abstract: Numerical prediction is one of the important ways to forecast gust. This study analyzed and evaluated the performance of the three gust diagnostic schemes (AFWA, UPP and IUM) in the CMA-BJ for gale forecasting over Beijing. Subjective analysis on two extreme gale processes as well as objective validation on batch experiments of each season indicated that there are significant differences in the gust forecasting using the three schemes, and the forecasting with the IUM scheme shows significant advantages. The IUM scheme performs relatively well on forecasting gales associated with both cold air and thunderstorms. For the gale on 18 March 2020, which was induced by cold air invasion, the scheme shows good performance on the forecast of starting and ending time, the location and evolution, and the maximum value during the entire period of the gale. For the gale on 2 August 2020, which was generated by a thunderstorm, the scheme overestimates the gale area and shows certain biases in the location of the gale. However, the scheme issues warning for gale, which is the most significant. Moreover, the gust speed predicted by the IUM scheme has positive biases at all seasons, but the forecasts of gales equal to or greater than level 5 are more consistent with observations. In general, the IUM scheme performs better than the other two schemes for the gale forecasting over Beijing, and can provide strong support for the gale forecasting.