The main objective of this study was to identify and investigate the atmospheric dispersion of fine particles from a cement plant stack in Doroud (Iran). For that, we used the SCREEN3 software, developed by the Environmental Protection Agency, and a Gaussian plume model for particles dispersion from the source. Simulated particles concentrations in ambient air were compared with measured data to evaluate the performance of both models. Upwind-downwind method was also applied to evaluate the ground-level PM10, PM2.5 and PM1.0 concentrations in the cement industry surroundings. The maximum particle concentrations were obtained within a radius of 1600-1800 m around the source in downwind direction. Within this area the modelled concentrations were strongly correlated to in-situ measured data (r2 = 0.98). The simulated and measured results show that the PM10 concentrations have significantly exceeded the daily (50 μg/m3, about 62.8%) and annual (20 μg/m3, about 82.3%) limit values established by the World Health Organization. Therefore, due to hazardous emissions a distance greater than 7500 m, in downwind direction of prevailing wind, is recommended to build residents' habitation. Quantification of air pollution impacts on the public health has increasingly become a critical component in the policy discussion. Both models provide useful information to define pollution areas for the environmental and health impact assessment and to provide recommendations for the implementation of effective local policies and mitigation strategies for air pollution. © 2018 Elsevier Ltd. All rights reserved.