This article studies the infl uence of building street geometry, wind direction, daily car traffi c and pedestrian trajectories in short-term personal exposure to PM 10 on a street canyon in Barreiro City, Portugal. An automatic system for the analysis of traffi c profi les, a Gaussian model for the determination of traffi c emissions (ADMSUrban) and a computational fl uid dynamics model (ANSYS Fluent) to simulate the dispersion of pollutants inside a street canyon were used. Buildings height, width, length and geometry, as well as the distance between the buildings and road width were tested. Besides the real actual geometry of the street, with the real disposition and volumetric confi guration of buildings, three scenarios were evaluated: (i) gaps of 4 m between the buildings along the street; (ii) gaps of 6 m between the buildings along the street and (iii) same height and width for all buildings along the street without gap between buildings. Wind direction, wind velocity and four pedestrian trajectories were considered. The results show that PM 10 concentration, at 1.5-m high plane, is highly dependent on street geometry and wind conditions for along-canyon wind directions (west and east wind). Lower concentration levels are obtained for confi guration (iii) because this geometry promotes the dispersion of pollutants under along-canyon wind direction. For cross-canyon wind directions (north and south), confi guration (i) results in lower PM 10 concentrations. There are no visible improvements in having higher gaps between buildings (confi guration ii). Short-term pedestrian personal exposure is dependent on the pedestrian trajectory considered inside the canyon. Pedestrian trajectories that correspond to crossing the road in the centre of the street result in the highest values.