In addition to total CO2 emissions, an intensity indicator is also important to assess CO2 emissions from electricity generation, allowing the spatial differences to be analyzed, which is crucial for regionally heterogeneous countries like China. This study analyzed the driving forces of aggregate carbon intensity (ACI), a measure of CO2 emissions per unit of electricity generation, in China between 1995 and 2014, using the LMDI method incorporating geographical effects. Both the entire study period and individual 5-year periods were analyzed. The results show that the geographic distribution effect, utilization efficiency effect, and thermal power proportion effect were responsible for a decrease in ACI during this period, whereas the energy composition effect showed an inhibitory influence on reduction. The dominant factor was the utilization efficiency effect, and the thermal power proportion effect showed the largest growth. Geographic distribution is a non-negligible factor in the reduction of ACI, and the regional ACI can be reduced by redistributing electricity generation to areas with lower ACI through trans-regional policy guidance. For future policies on reduction of ACI from electricity generation, this study proposes focusing on the thermal power proportion and geographic distribution effects to the same extent as on the utilization efficiency effect.