In this work we apply a trajectory classification approach, aiming at identifying the role exerted by meteorology on air quality in the Naples urban area (Southern Italy). Towards that end, the HYSPLIT (HYbrid Single-Particle Lagrangian Integrated Trajectory) model was used; backward trajectories were calculated for the period 1995-2004 and classified, based on a combination of the k-means and PCA (Principal Component Analysis) approaches. The robustness of the classification was strengthened using an ensemble approach, based on back-trajectories calculated for different (in the horizontal and vertical directions) arrival points around Naples and different arrival time. Eight clusters were identified, and the effects on air quality were evaluated. We focused on ozone and PM10 concentration, and evaluated to what extent the prevailing meteorological conditions were associated with different levels and temporal profiles of these two pollutants. We found that ozone and PM10 profiles share some similarities since they both load high during anti-cyclonic, subsiding conditions, a common situation during the summer months. During these days stagnation and recirculation effects enhance the concentration of locally emitted air pollutant. We also found that ozone exhibits a marked seasonal cycle, whose maxima appears during the summer season, but the discrimination with respect to the area of origin indicated the presence of a second relative maxima, appearing early during the spring season. PM10 pollution levels often exceed the annual limit of 50 microg/m(3). The days associated with a significant exogenous contribution from the North African countries were identified.