Atmospheric environmental impact assessment is an important part of environmental impact assessment, and the most important part is to calculate the ground pollution concentration under various atmospheric diffusion modes, and then evaluate according to the results. To facilitate the prediction of regional environmental air quality, a set of regional environmental air quality model system software was developed using Matlab. Based on Matlab, a mathematical model system of the chemical plant area is built and simulated. This model constructs the cloud model's comment set, weight set and assessment matrix. The final assessment results are also represented by the cloud models. The randomness and fuzziness of the assessment process and assessment results are quantitatively described by entropy and super-entropy, which make the assessment more accurate, and also extend the representation method of the assessment results from points to intervals, so that the description of randomness and fuzziness of the assessment results is more refined, and the model can better represent the overall condition of air quality. The proposed method quantitatively describes the fuzziness and randomness of the assessment process and the assessment results, and extends the representation method of the assessment results from points to intervals, which makes the description of the results more refined and can better represent the overall condition of air quality.