The spatial distribution of air pollution on the regional scale (Central Bohemian basin) is simulated by means of Charles University Lagrangian puff model SMOG nested in NWP model ETA. The results are used for the assessment of the concentration fields of ozone, nitrogen oxides and other ozone precursors. Current improved version of the model covers up to 16 groups of basic compounds and it is based on trajectory computation and puff interaction both by means of Gaussian diffusion, mixing and chemical reactions of basic species. Results of summer photochemical smog episode simulation are compared to results obtained by another couple adopted in framework of the national project as a basis for further development of data assimilation technique, Eulerian CTM CAMx nested in NWP model MM5. For the study being presented quite huge database of real emission sources was incorporated with many kinds of anthropogenic sources included in case of SMOG-ETA couple. There are some difficulties with the accommodation of the same database for use in both types of models described and their possible impact on the results are discussed as well as advantages of individual methods. Differencies in air pollution parameters due to the impact of contrast between real mesoscale meteorology of MM5 in resolution about 3 km and rather general representation of meteorology in lower resolution of NWP regimes of both models are discussed as well.