Beginning May 2005, the Institute of Multiscale Air Quality Studies (IMAQS) of University of Houston has been operating an air quality forecasting system (AQFS) to give a two-day ozone and particulate matter (PM) forecast for Southeast Texas (SETX). This paper evaluates the performance of the AQFS during the period of June 2005 to May 2006. The focus of this study is the performance of ozone forecasting and its spatial, seasonal as well as diurnal patterns at the model's innermost domain which covers SETX. The key findings are: 1) the system is capable of tracking day-to-day ozone variations in the area 2) the seasonal cycle is captured well by the model, with a double peak in late Spring and Fall, and a dip in July 3) the model showed better skill in predicting averaged (either spatially, temporally or both) ozone 4) most of model errors can be traced back to the imperfect meteorology from MM5 5) while overall model's day time peak matches observation, the night time forecasts showed substantial overpredicting which probably are caused by model's strong night time wind.