Abstract: In order to evaluate the forecast performance of CMA models in heavy rainfall processes associated with southwest vortex in Sichuan, using the object-based verification, the capabilities of the CMA-MESO and CMA-GFS models in predicting different types of warm-sector heavy rainfall events related to southwest vortex are compared and analyzed. The results show that: (1) For the Jiulong vortex-type warm-sector heavy rainfall events, the CMA-MESO model shows more stable spatial precipitation forecasting capability, offering higher reference value than the CMA-GFS model. (2) For the basin vortex-type warm-sector heavy rainfall events, although the CMA-GFS model shows slightly more stable spatial precipitation forecasting than CMA-MESO, the overall spatial verification score of CMA-MESO is higher. (3) For different types of warm-sector heavy rainfall events associated with southwest vortex, the CMA-GFS model forecasts the location and orientation of rainbands for heavy rainfall (≥25 mm) in relatively good agreement with observations. However, it underestimates the mean precipitation, peak intensity, and coverage area for the Jiulong vortex-type events. (4) For most periods, the CMA-MESO model generally forecasts the location, mean intensity, and peak intensity of heavy rainfall (≥25 mm) in close agreement with observations, though it tends to underestimate the coverage area. (5) The two models have the best prediction performance for the height field, followed by the wind field, while their performance in forecasting low-level temperature and humidity fields is relatively poor.