Abstract: To compensate for the disadvantages of the conventional verification methods and find the potential value of forecasts, a new verification method is developed for severe convection forecasting and nowcasting. This method is based on the characteristics of severe convective weather, the requirements of the very short-term forecasting and nowcasting operations, and performance evaluation of user surveys. In the approach of the method, forecasts for severe convection are compared to the observations aiming at their differences in intensity, area, distance, morphology and similarity. Several verification methods have been adopted for grid-based, site-based, and probabilistic forecasts. A forecast adjustment and integration method is also developed based on preference and tendency of users, and some verification indices and standards have been established. A series of operational verification products of this new method are developed for the national severe convection forecasting operations. This paper describes the design of this new method and its algorithms, as well as its application in Shanghai.