The contamination caused by several metals, among them mercury causes several health problems due to its bioaccumulation, so it was necessary to build models to determine it concentration. Some multimedia models have been created to determine metal concentration over time. In this study. Mercury was analyzed in its inorganic (HgCl2) and organic (CH3HgCl) form. They will be built in two different environmental compartments, water and sediment, in Sepetiba Bay, a water body located near the city of Rio de Janeiro. For the construction of this, adaptations were necessary in other models already know in the literature defined for lakes, taking to account that the bay would behave in a similar way, in addition to other considerations such as the type of existing emissions. In its construction, it was necessary to find data of direct emissions of this contaminant in the bay, outflows of water and air from that region, a dimension of each of the segments on this location, in the case of data were not found specifically forthe Sepetiba Bay, in addition for rainfall and aerosols input data, values from other similar works from other bay were also considered, from wherevolatility coefficients were removed between the segments and the portioning of these contaminants between the media. Knowing the local factors and building the equations provide by the modeling , it was possible to calculate, using Python 3 as a tool, the behavior of these contaminants over time, and compare the values obtained by these models with data provided by the literature and verify the discrepancy between them, that for the case of the selected time period, 2002 to 2018, the concentration obtained experimentally was 75.70*10-3 mg of total mercury per kg of contaminant and the one found through the constructed mathematical model was 74.39*10-3 mg of total mercury per kg of contaminant, thus having a difference of less than 41 percent in mass from the experimental values to the mathematically modeled values.