In the practical application of air protection, diverse dispersion models are used to calculate the concentration of contaminants in the air. They usually involve a universal character, which typically makes them sufficient for use in almost all conditions, with the exception of those clearly deviating from the average. This is especially relevant to industrial objects of large areas, introducing a great amount of heat and mechanical energy into the air. For such cases, the standard models can be extended in order to adapt them to the unusual local diffusion conditions. Next, to be applied in practice, they must have undergone validation to document the correctness of its operation. The article describes the process of validation of the air quality assessment model containing extended procedures to incorporate special factors affecting atmospheric dispersion in a coke industry. The set of statistical indicators, obtained on the basis of SF6 field experiment, evaluate its performance. The short comparison with some popular models of general-purpose character and an assessment of the suitability of individual indicators for validation purposes are also presented.