Global warming has become a serious challenge that threatens human sustainable development in the 21st century. In the context of climate change, reducing greenhouse gas emission is important. In this study, we take 98 countries worldwide as the case study area and calculate the spatial interaction degree generated from the total merchandise and fossil energy trades through the gravity model. We then apply the spatial econometric model to explore the driving mechanism and spatial spillover effect in 2000, 2005, 2010, and 2014. We further explore the similarities and differences with respect to the driving mechanism of carbon emission between developed and underdeveloped countries. Results indicate that: (1) The spatial spillover effect increased in 2000, 2005, 2010, and 2014. (2) Urbanization, per capita GDP, and energy intensity had significant positive influences on per capita CO₂ emissions, and the renewable energy use rate negatively influenced per capita CO₂ emissions. (3) The contributions of urbanization, per capita GDP, and energy intensity were more powerful in underdeveloped countries than those in developed countries, whereas the use of renewable energy had a greater influence in developed countries than in underdeveloped countries. Accordingly, developed and underdeveloped countries should strengthen their cooperation and exchange of low-carbon technology. The energy structure should also be optimized to reduce CO₂ emissions.