Abstract: To fully understand the impact of assimilating ground-based aerosol observations on the forecast performance of the CMA Regional Chemical Weather Numerical Prediction System (CMA-CW v1.0) rapid cycle system, continuous cycle tests were conducted over a one-month period in both winter and summer at 6-hour intervals. Results of 72 h forecasts at 00: 00 UTC (Coordinated Universal Time) were examined and evaluated, and the conclusions are as follows. (1) The CMA-CW v1.0 rapid update cycle system is capable of efficiently assimilating ground-based aerosol observation data. The December 2023 experiment significantly improves the accuracy of PM_2.5 mass concentration analysis, reducing the monthly mean root-mean-square error by approximately 25 μg/m³, representing a relative reduction of about 47%; It also significantly improved the accuracy of 24-hour forecasts for PM_2.5 mass concentration and visibility, with a marked reduction in instances of missing aerosol pollutant concentration forecasts. (2) The assimilation of ground-based aerosol observation data not only improved the forecasting capability for pollutant concentrations but also made a significant positive contribution to the forecasting of weather elements; the winter experiment (December 2023) indicated that aerosol data assimilation made a positive contribution to light rain forecasts, while the summer precipitation experiment (July 2025) showed that aerosol assimilation made a positive contribution to 24-hour precipitation forecasts across all intensity levels; however, aerosol observation assimilation caused a slight warming in the 2-meter temperature forecast but had no impact on the 10-meter wind forecast. Ground-based aerosol assimilation significantly improves the forecast quality of PM_2.5 mass concentration and visibility, and also makes a positive contribution to precipitation forecasts. The CMA-CW v1.0 rapid-cycle system has been put into operational use and can provide reliable technical support for air quality forecasting and early warning.