Neural network model of distribution of pollutant in the atmosphere of the city of Taganrog has been developed on the basis of monitoring data, meteorological information and calculation made by software «Ecolog-gorod». The results of the work of model have been visualized in software ArcGis. Distribution maps of nitrogen dioxide has been built under the conditions of eight directions of wind, two wind speeds and two temperature ranges.

Keywords: Neural network, prediction model, urban air pollution