Abstract:In order to explore the distribution of air pollutants in different regions and at different times in the northern Anhui urban agglomeration and the relevant variables affecting the concentration of PM2.5 in the air, combined with the statistical data of Anhui Provincial Department of Ecology and Environment from 2018 to 2021 and relevant regional data, The principal component analysis (PCA) method was used to analyze the correlation between the monthly data and the annual data on the influencing factors of air quality respectively according to the time period and seasonal changes. The changes in the concentration of air pollutants PM2.5 and PM10 and other air pollutants in different seasons were compared, and the backpropagation neural network (BP) based on PCA algorithm was constructed. PCA-BP model was established and cross-validation method was used to improve the accuracy of the model, and short-term prediction of PM2.5 concentration in the atmosphere was made. The experimental results show that the main influencing factors of PM2.5 concentration are PM10, CO, NO2 and SO2. PM2.5 content in northern Anhui is higher in winter. The accuracy of the prediction model was higher in summer and autumn, but lower in winter. The prediction accuracy R2 of the four seasons reached 0.924, 0.958, 0.935 and 0.794, respectively.