Abstract. The resolution of regional chemical/dynamical models has important effects on the calculation of the distributions of air pollutants in urban areas. In this study, the sensitivity of air pollutants and photochemical ozone production to different model resolutions is assessed by applying a regional chemical/dynamical model (version 3 of Weather Research and Forecasting Chemical model – WRF-Chemv3) to the case of Mexico City. The model results with 3, 6, 12, and 24 km resolutions are compared to local surface measurements of CO, NO_x, and O₃. The study shows that the model resolutions of 3 and 6 km provide reasonable simulations of surface CO, NO_x, and O₃ concentrations and of diurnal variations. The model tends to underestimate the measurements when the resolution is reduced to 12 km or less. The calculated surface CO, NO_x, and O₃ concentrations at 24 km resolution are significantly lower than measured values. This study suggests that the ratio of the city size to the threshold resolution is 6 to 1, and that this ratio can be considered as a test value in other urban areas for model resolution setting. There are three major factors related to the effects of model resolution on the calculations of O₃ and O₃ precursors, including; (1) the calculated meteorological conditions, (2) the spatial distribution for the emissions of ozone precursors, and (3) the non-linearity in the photochemical ozone production. Model studies suggest that, for the calculations of O₃ and O₃ precursors, spatial resolutions (resulting from different meteorological condition and transport processes) have larger impacts than the effect of the resolution associated with emission inventories. The model shows that, with coarse resolution of emission inventory (24 km) and high resolution for meteorological conditions (6 km), the calculated CO and O₃ are considerably improved compared to the results obtained with coarse resolution for both emission inventory and meteorological conditions (24 km). The resolution of the surface emissions has important effects on the calculated concentration fields, but the effects are smaller than those resulting from the model resolution. This study also suggests that the effect of model resolution on O₃ precursors leads to important impacts on the photochemical formation of ozone. This results directly from the non-linear relationship between O₃ formation and O₃ precursor concentrations. Finally, this study suggests that, considering the balance between model performance and required computation time on current computers, the 6 km resolution is an optimal resolution for the calculation of ozone and its precursors in urban areas like Mexico City.