In order to solve the issue that monitoring PM_2.5 levels based on sparse ground stations is insufficient to provide PM_2.5 concentration data with broad coverage and high spatial resolutions and support regional fine particulate matter pollution prevention and control,the province of Hubei was selected as an area of research.MODIS AOD was used as the main predictor and was combined with meteorological parameters,such as temperature,relative humidity,wind speed,pressure,and normalized difference vegetation index(NDVI) data,as auxiliary predictors in a linear mixed effects model to establish the daily relationship of AOD-PM_2.5 from 2015 to 2017 in order to estimate PM_2.5 levels in the study area.The model was validated by a tenfold cross validation (CV) method.The results showed that the model performances were quite satisfactory,with cross-validated correlation coefficient values (CV R²) between the PM_2.5 estimations and observations of 2015-2017 of 0.89,0.85,and 0.88,respectively,suggesting that this method can be used to monitor regional PM_2.5 concentrations with rather high accuracy.The temporal and spatial variation characteristics were analyzed based on the observed and model-estimated PM_2.5 data and showed that distinct spatial variations of PM_2.5 concentrations existed,with high values in the east,south,and north and low values in the northwest and southeast.In addition,a declining trend of PM_2.5 concentrations in Hubei province was observed,with the annual mean values for 2015-2017 of 65.6±39.8 μg/m³,57.1±34.1 μg/m³,and 48.1±28.3 μg/m³,respectively.In terms of city,all cities showed a downward trend,except for Xianning and Suizhou,which had steady annual mean values in 2016 and 2017.