Utilizing the Anomaly Numerical-correction with Observations（ANO）based on historical observation data and anomaly integration，in conjunction with the ERA5 reanalysis data，a rectification test was conduct‐ ed on the forecasting products of the Southwest Center WRF-ADAS Real-time Modeling System（SWMS in short）. This study evaluated the efficiency of the ANO method in enhancing short- to medium-term weather fore‐ casts for meteorological quantities during a catastrophic regional heavy rainfall event over the complex topography from June 20 to 25，2019. Results revealed that the SWMS model exhibited commendable predictability in the middle and upper troposphere，although its accuracy gradually reduced in the lower layers. After post-processing correction using the ANO method，all the predicted variables showed obvious improvements. The aver‐ age Anomaly Correlation Coefficient（ACC）for the 500 hPa and 700 hPa geopotential height fields within the 72-hour integration increased by a range of 0. 1 to 0. 2，reaching approximately 0. 8，while the 850 hPa geopotential height ACC showed a maximum enhancement of 0. 6. Concurrently，the Root Mean Square Error （RMSE）for the corrected 700 hPa and 850 hPa geopotential heights exhibited significant reductions with an average decrease of 24% and 66%，respectively. The correction outcomes for temperature，specific humidity，and horizontal wind also displayed positive effects，which reveal the efficient correcting performance of the ANO method based on historical observation data in rectifying short- to medium-term numerical forecasts of the SWMS model over complex topography.