Weather Research and Forecasting (WRF) model coupled with Community Multiscale Air Quality (CMAQ) model was used to simulate temporal and spatial distribution and multiscale transport patterns of PM2.5 among 2 large regions (NORTH and SOUTH), 8 small regions (North, Northeast, East, Center, South, Southwest, Northwest and others) and 31 provinces in China in January and July of 2015. The simulated PM2.5 concentrations were compared with the observed values, showing the accuracy of WRF-CMAQ. In January and July, local emissions were the major contributors to PM2.5 concentrations with the percentage exceeding 60% in seven regions in China. Among 31 provinces, external transport was dominant source of PM2.5 in some provinces, such as Shanghai and Qinghai, and local emissions were the main contributors in other provinces, such as Hebei and Xinjiang. We also identified the major contributor of PM2.5 in each region, indicating distinctly different regional transport patterns among provinces and regions in China.