The understanding of distribution of gases such as ammonia (NH3) in agricultural installations is of growing importance due to its effect on health and productivity of animals and workers. There are methodologies available for determination of NH3 emissions in poultry houses by continuous monitoring like the portable monitoring unit (PMU) and mobile air emissions monitoring unit (MAEMU) methods, tracer gas ratio method and Model-based approach of mass balance. However all methods require long experimentation periods. Computational Fluid Dynamics (CFD) is a powerful and efficient tool which allows for prediction of this distribution of gases in real time, which allows for a reduction of the number of experiments. Based on these facts, the aim of this study was to use CFD to develop and validate a model to determine NH3 emissions in a non-insulated broiler house installation with natural ventilation, typical to subtropical and tropical countries. It was found that the proposed model showed a good statistical correlation with the experimental data that were based in the methodologies used by Wheeler et al. (2006) and Osorio (2011). Therefore this CFD model could be used to predict behavior of NH3 emissions in real time generated by poultry facilities.