Abstract: The performance of the CUACE system for spring air quality forecasts in Shenyang was verified and modified using the atmospheric environment observational data during April and May 2015.The results indicate that its forecasting for the mass concentrations of six pollutants,i.e.,PM₁₀,PM_2.5,NO₂,O₃,SO₂,and CO,is generally smaller than the observed values,especially for the PM₁₀ concentration that is significantly underestimated.The PM_2.5 is daily primary pollutant in spring in Shenyang predicted by the CUACE system,while the observations show that there is comparable number of days when PM₁₀ and PM_2.5 both are the primary pollutants.Meanwhile,there is a large difference in the air pollution grades forecasted by the CUACE system and determined by the observations.The results forecasted by the CUACE system are modified using a linear fitting formula based on the forecasting and observed pollutants' concentrations.Through this modification,the forecasting results on the primary pollutants are principally consistent with the observed ones.Furthermore,the forecasting accuracy on the air pollution grade is improved by about 50% to 80%.