摘要： Visible Infrared Imaging Radiometer Suite (VIIRS) is a next-generation polar-orbiting operational environmental sensor, and its aerosol optical depth (AOD) data has been widely applied to ground-level PM2.5 prediction and air pollution research. However, VIIRS performance affects the accuracy of prediction. In this study, three types of VIIRS AOD products were collected from January 2014 to December 2016 to evaluate their performance under different air quality conditions. Near-surface PM2.5 concentrations were determined for the same period to categorize air quality as clean, moderate, or heavily polluted. The performance of three VIIRS AOD products was evaluated from the successful retrieval rate and AOD accuracy aspects. For clean weather days, the AODs obtained from VIIRS intermediate products (IP) had the lowest average absolute bias (0.15 ± 0.15). For the moderate and heavy pollution days, the average absolute deviations of VIIRS environmental data record (EDR) AOD (Quality Flags = 3 and > 1) products were lowest at 0.14 ± 0.14 and 0.2 ± 0.1 respectively. These results suggest that the EDR AOD, Quality Flags > 1 and = 3, products were more suitable for heavy and moderate air pollution, respectively. During clean weather, performance of _IP AODs was found to be best. Moreover, seasonal analysis indicated that the EDR AOD (Quality Flags = 3) products were more suitable for spring and autumn and the IP AOD for summer. However, the performance of the EDR AOD products (Quality flags > 1) were best during winter. To ensure accuracy of PM2.5 predictions over central China, researchers should carefully balance the needs for successful retrieval rate and accuracy of VIIRS AOD products.