<p>Air pollution remains one of the most serious environmental problems in megacities, especially in cities located in complex geographical conditions. This study aims to investigate the spread of pollutants such as NO₂, O₃ and N₂O₅ in the urban environment of Almaty using numerical modeling based on computational fluid dynamics (CFD). To solve this problem, the Reynolds-averaged Navier–Stokes equations and the impurity transport equation were used. The finite volume method was used to discretize the corresponding equations. The validation between the experimental data and the numerical results obtained in solving the test problem using the k–w SST turbulent model confirms the correctness of the mathematical model. A sensitivity analysis of the computational grid was performed. The effect of different heights of the grass barrier on the concentration distribution and on the product of the chemical reaction was also considered. Three scenarios are considered in the work: no vegetation, use of the first type of trees and use of the second type of trees. The modeling results showed that both types of tree barriers have a positive effect on reducing NO₂, O₃ and N₂O₅ concentrations in the ground layers of the atmosphere, with the second type of tree demonstrating the greatest efficiency. This is due to the optimal crown shape, which facilitates the dispersion and absorption of pollutants. In the scenario using the second type of trees, the concentration of N₂O₅ was reduced by up to 8 times compared to the absence of vegetation. The obtained results emphasize the importance of integrating green spaces into urban planning as one of the key tools for reducing air pollution.</p>