In the new era of promoting ecological civilization construction, the in-depth implementation of the carbon peaking and carbon neutrality goals have become a hot topic of concern for all sectors of society, especially in the Yangtze River Delta urban agglomeration, which serves as a model for high-quality development of the national regional economy. So it is important to study the spatial distribution of carbon emissions and the multi-scale spatial relationships between carbon emissions and influencing factors. This paper selects the Yangtze River Delta urban agglomeration as research object and the county as the research scale, and applies spatial analysis methods such as Moran's I, cold and hot spot analysis to explore the spatial distribution of carbon emissions. At the same time based on the MGWR model analyzes the spatial variation relationship of carbon emissions and the influencing factors. The conclusion is as follows: (1) There are significant H-H (high-high) clustering and L-L (low-low) clustering of carbon emissions in the Yangtze River Delta urban agglomeration; (2) Compared to OLS model and GWR model, the results of MGWR model are more precise and smaller errors. They can reflect the impact of scale variation between different influencing factors and carbon emissions and ensure the results more accurately and reliably; (3) GPP, road density, GDP, the proportion of the primary, secondary and tertiary industry are the influence mechanism on the global scale. And NDVI, population density, and electricity consumption are the influence mechanisms at the local scale. According to the research results of this paper, emission reduction policies can be formulated based on spatial changes and global or local measures can be formulated based on bandwidth changes between carbon emissions and influencing factors to promote the achievement of the carbon peaking and carbon neutrality goals in the Yangtze River Delta urban agglomeration.