European Community Directive 2008/50/EC attributes a higher role to air quality models in air quality assessment and management, and establish criteria for acceptable model uncertainties or quality objectives as a function of pollutant. This study analyses the capacity of MINNI (Italian Integrated Assessment Modelling System for supporting the International Negotiation Process on Air Pollution and assessing Air Quality Policies at national/local level) atmospheric modelling system to simulate ozone concentrations over Italy for the years 1999, 2003 and 2005, according to the new EU criteria. Surface ozone concentrations over Italy exceed often the thresholds established in EU legislation to protect human health and to prevent damage to ecosystems. The yearly simulations were carried out using the national emission inventories for Italy and EMEP inventories for the other countries included in the computational domain. The meteorological conditions were simulated with RAMS driven by ECMWF initial and boundary conditions. The ozone formation and destruction under different conditions of solar radiation, temperature and chemical composition of atmosphere was computed with the photochemical mechanism SAPRC90. At most of the background stations (rural, Urban, subUrban) considered in the evaluation, the comparison of simulated and measured ozone concentrations shows a fair agreement between hourly observed and modelled ozone concentrations; furthermore RDE (Relative Directive Error) has values lower than that indicated in the directive. The modelling system exhibits good performance also with respect to other statistical indicators, such as MNBE (mean normalised bias error) which has values laying inside the range proposed in USEPA's guidelines for an acceptable level of air quality model performance.