As the world’s top carbon-emitting country, China has placed great emphasis on understanding the driving factors of carbon emissions and developing appropriate emissions reduction policies. Due to the obvious variations in carbon emissions among various industries in China, corresponding policies need to be formulated for different industries. Through data envelopment analysis, this study introduced the Shephard distance function into the logarithmic mean Divisia index (LMDI) for decomposition analysis, built a carbon emissions decomposition model of 23 industries in China during 2003–2015, and analyzed the impact of 10 factors driving carbon emissions. The main results are as follows. (1) Potential gross domestic production (GDP) is a crucial factor for increasing carbon emissions, whereas potential energy intensity and technological advances of carbon emissions have a significant inhibitory effect on carbon emissions; (2) the technological progress of energy usage and the technological advances of GDP output are manifested by inhibiting carbon emissions at the early stage of development and increasing emissions at the later stage; (3) the structure of coal-based energy consumption is difficult to change in the long term, resulting in a weak effect of energy mix on carbon emissions and an increase in carbon emissions due to the potential energy carbon intensity factor.

Against the backgrounds of emission reduction targets promised by China, it is crucial to explore drivers of CO2 emissions comprehensively for policy making. In this study, Shandong Province in China is taken as an example to investigate drivers in carbon density by using an extended Kaya identity and a logarithmic mean Divisia index model (LMDI) with two layers. It is concluded that there are eight positive driving factors of carbon density during 2000–2015, including traffic congestion, land urbanization, etc., and seven negative driving factors comprising energy intensity, economic structure, etc. Among these factors, economic growth and energy intensity are the main positive and negative driving factor, respectively. The contribution rate of traffic congestion and land urbanization is gradually increasing. Meanwhile, 15 driving factors are divided into five categories. Economic effect and urbanization effect are the primary positive drivers. Contrarily, energy intensity effect, structural effect, and scale effect contribute negative effects to the changes in carbon density. In the four stages, the contribution of urbanization to carbon density is inverted U. Overall, the results and suggestions can give support to decision maker to draw up relevant government policy.