We calculated energy-related carbon emissions from Beijing's production and household sectors in 1995, 2000, 2005, and 2009 based on the default carbon-emission coefficients provided by the Intergovernmental Panel on Climate Change (IPCC). Taking 1995–2000, 2000–2005, and 2005–2009 as the study periods, we decomposed the effects of changes in carbon emissions resulting from eight causal factors using the logarithmic mean divisia index method: economic activity and population size, which reflect the change in socioeconomic activity; energy intensity and energy consumption per capita, which reflect the change in the intensity of energy use; and economic structure, the urban and rural population distribution structure, and the energy mix of the production and household sectors, which reflect structural changes. We found that in all three study periods, the changes in economic activity, population size, and energy consumption per capita stimulated emissions, whereas the changes in energy intensity, the urban and rural population distribution structure, and the energy mix of the production and household sectors decreased emissions. In contrast, changes in the economic structure decreased emissions in the first and third periods, and increased emissions in the second period. Our results clearly indicate that under current practices, carbon emissions will increase as a result of rapid growth of the economy, the population, and the energy consumption per capita. In the future, the main routes to reduce carbon emissions will be to adjust the economic structure and energy mix, and to reduce the energy intensity of each sector.