Ambient air quality is inseparable from people's production and life and social development, so the atmospheric environment monitoring system plays an irreplaceable role in environmental protection and pollution control. Use the real-time updated monitoring information to grasp the air pollution situation, evaluate and predict the environmental quality, and provide technical support for early warning decision-making, scientific air quality control, and regional joint prevention and control work. The existing air quality numerical prediction system WRF-CMAQ, due to the uncertainty of the pollution source emission inventory and the inability to fully quantify the physical and chemical changes in the atmospheric transmission, leads to deviations in the air quality forecast values. In view of the current situation of the existing air quality forecasting system, a deep belief network (DBN) model is proposed to mine the relationship between the predicted value of the numerical forecasting model and the measured value in the area, and establish an air quality forecasting model based on deep learning. The model utilizes the historical monitoring data of multiple Chinese monitoring stations in the study area and the corresponding meteorological forecast data, and fully considers the temporal variation and spatial distribution characteristics of atmospheric pollutant concentrations. During the forecast period, the forecast value of pollutant concentration at any site in the region is corrected to improve the effectiveness of the air quality forecast model.