This study evaluates the performances of three Community Multiscale Air Quality (CMAQ) model v5.0 simulations that utilise meteorological input data from the 5th-generation Mesoscale Model (MM5), the Weather Research and Forecasting (WRF) model and WRF with Four-Dimensional Data Assimilation (WRF/FDDA). Three sets of CMAQ model simulations were performed for a high air pollution episode from October 9 to October 12, 2004 over the Pearl River Delta (PRD) region, which was caused by a high-pressure system. The observations at Guangzhou urban site (113.26°E, 23.13°N) were used to evaluate the results of meteorological models and air quality models. In all three experiments, the simulated pollutants were distributed mainly in the western part of PRD, but there was still a difference of horizontal distribution among them. The predicted wind fields, relative humidity and planetary boundary layer (PBL) height in meteorological simulations were the major factors contributing to this difference.