The work deals with the application of artificial neural networks (ANN) to the estimation of the surface distribution of chemical elements in the soil. For the study, a square area with a side of 1 m was chosen far from the sources of pollution. In this area, 100 cores of topsoil (0.05 m deep) were sampled. The specimens were analysed by the X-ray fluorescence spectrometer Innov X 5000. The best ANN structure for estimation of the surface distribution of chemical elements in the soil was selected after computer simulation. Comparison of concentration values of the chemical elements surface distribution in the soil made by the ANN with known values showed that a trained ANN gives prediction models comparable in accuracy with other methods as an interpolator and as the forecast method, as well.