In the last years, the fast grouth of the urban centers has shown a need to study how human action in water, soil and on the air affects natural resources. A future lack of water in the next decades has been observed by many studies and much effort has been devoted to find strategies which will help to manage properly water resources. It is necessary to develop new experimental techniques and numerical models with a view to properly evaluate and hence minimize pollution emission. In this scenario, it is of paramount importance to evaluate what will be the environmental impact of new effluent emissions in rivers or bodies of water. An in-house CFD model to predict the dispersion of effluents (or pollutants) along a river has been developed. The most important parameter for this model is the effluent dispersion coefficient in the river section (from the mass conservation equation). This coefficient is dependent on several factors, such as: river shape, flow characteristics, effluent dispersion coefficient, turbulence level, among other factors. There is a lack of explanation in the literature about the proper use of these coefficients and the danger of using the same value for the dispersion coefficient for models based on different flow assumptions, especially if the model uses different assumption for the river shape. This work discusses these aspects and presents a methodology to determine the best value for the coefficient for a given model. The methodology is presented for the model of this work. The dispersion coefficient prediction is based on experimental data and a statistical model especially developed for this model determines the best dispersion coefficient for the experimental data collected. A real case based on an evaluation of the environmental impact caused by the emissions of an industry based in Campinas, Sao Paulo state, Brazil is presented. Results indicate the software is suitable for the prediction of multiple effluent emissions in rivers.