The data of Global/Regional Assimilation and Prediction SystemRegional Ensemble Prediction System (GRAPESREPS) from June to August 2015 are used and an objective method based on grid data is designed for locating the center position of the southwest vortex. Firstly, the accuracy of GRAPES to the southwest vortex is tested, then 4 longlife southwest vortexes are picked to evaluate the ensemble forecasting performance of GRAPESREPS in the generating, developing, moving and precipitation process of southwest vortex. The results show that: (1) accuracy of GRAPES to the southwest vortex keeps a high score, and the false rate is slightly larger than the missing rate. (2) GRAPESREPS performs well in forecasting generation and development of the southwest vortex. Most ensemble members can predict these processes, but the occurring time is earlier than observation. (3) The forecasted moving path for southwest vortex by GRAPESREPS in 24 h is reasonable, the ensemble forecast is averagely much better than control member, and the path goes to north after 24 h in the eastward moving cases. (4) The forecast of the intensity of the southwest vortex is too strong, with higher value of center positive vorticity and lower value of geopotential height. (5) In 24 h forecast, the rain score for light rain to heavy rain triggered by the southwest vortex performs well, and the rainfall area matches the observation very well, while area for torrential rain lies slightly north in some cases, but mostly is reasonable. After 24 h, since the pathway of the eastward moving southwest vortexes is by north, the prediction of strong rainfall area follows the same trend. It can be seen that the system has the ability to forecast precipitation caused by the southwest vortex. So, improving the capability of mesoscale ensemble forecast by GRAPESREPS would be helpful for the prediction of precipitation caused by the southwest vortex.