Since air pollution impacts our health and the environment, air quality forecasts represent information of primary importance, especially for an area susceptible to episodes of pollution such as the Po Valley. The Environmental Protection Agency of the Italian region of Veneto ARPAV) has implemented a system based on the CAMx photochemical model to predict concentrations of particulate matter and ozone. Meteorological input is estimated on the basis of the numerical output from the COSMO limited area model. Emissions are calculated using the INEMAR software and the concentrations at the boundaries of the computational domain are provided by the PREV’AIR modelling system. The system generates, on a daily basis, forecasts of up to three days and produces interactive maps which can be viewed from the agency’s website. An evaluation of the forecasts of 1hr-ozone daily maximums and PM10 daily means has been conducted, by means of a statistical comparison with a set of air quality stations. The result is satisfactory since the forecasts comply with the benchmarks established for the model evaluation. The forecast skill respect to the persistence method has also been investigated.