This paper evaluated dispersion of PM2.5 emitted from roasted chicken restaurants around Metropolitan Area of Huancayo city, Peru. It was used topographic data with 30 m of resolution, surface and vertical structure meteorological data. Main characteristics from emission sources also were used as input in the AERMOD model. Then for evaluating the model efficiency, as reference it was used data from particulate matter samplers at three urban sites of Metropolitan Huancayo (named HYO, UNCP and CHI). The results showed the association of the wind speed and direction with the predominant dispersion of PM2.5. Differences among the sites were registered but CHI showed highest values. The concentrations from AERMOD model were lower than sampled ones, mainly because model only was focused on restaurants meanwhile samplers collected particulate matter from all other emitters. AERMOD provided good options to model air pollution dispersion at mountains regions for air quality management.