Work is devoted to the use of artificial neural networks (ANN) to restore the surface distribution of chemical elements in the soil. For the study was chosen square area with a side of 1m in the Yamal-Nenets Autonomous Okrug, Tyumen region far from the sources of pollution, where has been selected 100 samples of topsoil. Samples were analyzed by X-ray fluorescence spectrometer Innov XX 5000. Studied in detail and identified the best structure of artificial neural networks to restore the surface distribution of chemical elements in the soil. Different approaches to providing data for analysis by artificial neural networks was tested. It is shown that the presentation of data on the concentrations of elements in the form of geochemical spectrum is perspective for ANN method.