After processings of installation, adjustment, localization and automation, the 3rd v ersion of city air pollution prediction system (CAPPS3) is used to establis h the regional air quality numerical forecast system suited to Liaoning to forca st city pollution index, pollution level and the daily mean concentration distri bution of pollutants in Liaoning. Comparisons between forecast results and obser vation data show that the forecast accuracy of SO2 and NO2 in the 14 main cities of Liaoning are higher than th at of PM10. The location and value of regional pollutant concentration ar e realistic and the change of hourly pollutant concentration is also consistent with the ch ange of real wind, indicating that the CAPPS3 is able to exactly forecast the genesis and develo pment of regional pollution in time.