This paper suggests an approach to process and model the data obtained in biomonitoring studies. The approach is validated on data obtained from biomonitoring studies performed in the Republic of Moldova in 2015. Using the preliminary data, the decomposition on the basis of the pollution spectrum for the most polluted and cleanest sites is first carried out. The deviations of model predictions from the actual measurements are considered. A correlation analysis is next performed to evidence the correlation of two geographical coordinates with chemical elements. Factor analysis and regression analysis are applied to highlight the nonlinear mechanisms specific to the obtained data. A multilayer neural network-based model is derived to describe the relationship of the pollution rank to the geographic coordinates. The predictive capabilities of the model are represented graphically.