In order to study the relationship between precipitation and atmospheric pollutants, according to the air quality data of 113 key environmental protection cities in China in 2017, the spatial distribution law of urban clusters and the linear fitting equation between them are obtained. Based on the distribution characteristics of precipitation in the north and south and the urban air pollution index, the fuzzy clustering method was used to classify cities with different degrees of affinity. The Spearman correlation and fitting regression method were used to analyze the air pollution characteristic value and precipitation. The results show that when the key environmental protection cities are clustered into five categories, the spatial distribution of "rings and loops" is presented. The innermost ring basically covers five provinces of Hebei, Henan, Shandong, Shanxi and Shaanxi. The rest of the city is distributed between the second and third rings, and the urban air quality gradually increases from the inside to the outside. As the number of clusters increases, the urban agglomeration pattern changes, and the spatial distribution develops into "distribution according to the basin", but the "circle" city does not change much. In this clustering situation, the impact of the amount of precipitation on the atmospheric quality environment of environmentally friendly cities can be clearly reflected. The precipitation is negatively correlated with the concentration of atmospheric pollutants, and the environment in these five provinces is seriously polluted. Using Kolmogorov-Smirnov test method, it is calculated that the overall distribution of NO2, O3 and PM2.5 conforms to the normal distribution, and that of SO2, PM10 and CO does not conform to the normal distribution, nor does it conform to the uniform and exponential distribution. The correlation coefficients of NO2, O3, PM2.5, SO2, PM10 and CO with precipitation are -0.316, -0.238, -0.332, -0.617, -0.574 and -0.695, respectively. O3 is in weak correlation; NO2 and PM2.5 are in moderate correlation; SO2, PM10 and CO are in strong correlation. The coefficient of linear regression equation is between -0.031 and -0.008, and the negative correlation between precipitation and atmospheric pollutant concentration is obvious. The research results will help to understand the impact of precipitation on urban atmospheric environment changes, and provide the necessary theoretical support for the sustainable development of urban construction.