Urban agglomerations are of utmost importance in facilitating the progress of urbanization and industrialization. The study of the factors influencing CO2 emissions, particularly when considering the cities’ specialized industrial economy structure within urban agglomerations, holds significant importance in developing effective strategies to reduce CO2 emissions. Using the cross-sectional dataset, this paper initially employed cluster analysis to classify the sample cities according to the proportion of output values. Subsequently, based on the obtained classification results, the driving forcesDriving forces of CO2 emissions at the municipal level were examined by utilizing quantile regressionQuantile regression. The findings indicate that: (1) The sample cities can be classified into five groups according to their pillar industries, with significant disparities in CO2 emissions among these five types of cities. (2) Population growth, the rise in the share of secondary industry, and land urbanization actually lead to an escalation of CO2 emissions. The latter two factors have a more pronounced impact on cities with both higher and lower CO2 emissions compared to cities with moderate emissions. (3) The adoption of a polycentric and highly connected development model can facilitate the reduction of CO2 emissions, particularly in areas with high levels of emissions. (4) The case of the YRDUAYangtze River Delta Urban Agglomeration" (YRDUA does not support the traditional Environmental Kuznets Curve hypothesis. Our findings offer a foundation for the low-carbon construction in urban agglomerations from the perspectives of socioeconomic progress and urban structure organization.