The accuracy of an environmental impact assessment study in analyzing the consequences of either continuous or accidental releases in the atmosphere is critical especially when adverse health effects are expected to be found. High accuracy is hard to be achieved in complex areas where land/sea breeze and topography give rise to complex circulation patterns. In these conditions stationary modeling systems such as the Gaussian often fail, providing misleading results which in turn lead to poor mitigation and control strategies. Conversely, advanced modeling systems, able to deal with inhomogeneous and non stationary atmospheric phenomena, can be expected to achieve better accuracy. The present study shows the application of a modeling system, composed by a diagnostic meteorological model and a Lagrangian particle model, to assess the environmental impact of a cement plant located close to a populated urban area in central Italy and characterized by complex topography and atmospheric circulation. 3-D wind and temperature fields, produced by interpolation of ground and vertical remote sensing meteorological data, have been provided to a particle dispersion model to calculate 3-D concentration of NOx, SO2 and CO in a 10 × 10 km2 domain. Comparison of results with observations confirms that the modeling system is able to achieve high accuracy when high quality data are provided as input. The applied methodology has shown that a modeling system can be used to provide local authorities and plant managers with a valid tool for consequences analysis studies.