To investigate the spatiotemporal patterns and agglomeration characteristics of carbon emissions in the Pearl River Basin, we constructed a carbon emission estimation model by coupling multi-source data. The spatiotemporal dynamics and spatial correlation characteristics of urban carbon emissions were explored using exploratory spatiotemporal data analysis and modified gravity modeling. The findings indicate that the total carbon emissions in the Pearl River Basin increased from 312.67 million tons to 336.54 million tons. Dongguan, Shenzhen, and Guangzhou consistently stood out as cities with the highest carbon emissions. On the grid scale, the high-value carbon emission agglomeration expands towards the periphery, with the Pearl River Delta region serving as the core, whereas the high-value carbon emission area in the middle and upper reaches is characterized by a point-like distribution. Carbon emissions in the Pearl River Basin show a positive spatial correlation, although there is a decreasing trend in the spatial interaction effect. Furthermore, there is a positive synergistic trend among neighboring cities in terms of carbon emissions. The average linkage intensity of urban carbon emissions increases from 5.93 to 18.97, indicating strengthened connectivity among cities. The carbon emissions network structure shows a trend towards centralization. This method incorporates carbon sources and sinks into the calculation process, has potential practical value, and can support the development of a carbon reduction strategy.