•This paper analyses the factors affecting CO2 emissions of OECD.•LMDI decomposition and decoupling analysis methods are combined.•We find that the impact of population distribution on CO2 emissions is negligible.•Energy intensity and per capita GDP are the main impact factors of CO2 emissions.•Technical factors have a greater impact on decoupling elasticity. Under the framework of the Kaya identity, this paper uses the Logarithmic Mean Divisia Index (LMDI1Abbreviation: LMDI, the Logarithmic Mean Divisia Index.1) decomposition method to explore the impacts of CO2 emission intensity of fossil energy, energy consumption structure, energy intensity, per capita Gross Domestic Product (GDP2Abbreviation: GDP, Gross Domestic Product.2), population distribution, and population size on CO2 emissions in the Organisation for Economic Co-operation and Development (OECD3Abbreviation: OECD, the Organisation for Economic Co-operation and Development.3) from 2001 to 2015. Additionally, the Tapio decoupling analysis is used to explore the decoupling relationships between the above influencing factors and CO2 emissions. Moreover, the LMDI decomposition formula is embedded into the decoupling analysis to analyze the influences of technical and non-technical factors on above decoupling elasticity. The results indicate that energy intensity and per capita GDP are the main factors affecting CO2 emissions. The former is the main reason for the decrease in CO2 emissions, and the latter is the main reason for the increase in CO2 emissions. The impact of population distribution on CO2 emissions is negligible. The decoupling states between the overall CO2 emission intensity of fossil energy, energy consumption structure, energy intensity, per capita GDP, and population size and CO2 emissions during 2001–2015 are recessive decoupling, recessive decoupling, weak negative decoupling, strong decoupling, and strong decoupling, respectively. Moreover, the influence of technical factors is greater than that of non-technical factors, and their influence directions are always opposite. In addition to our primary contributions, there are three marginal contributions in this paper. First, the population distribution is included in LMDI factorization. Second, LMDI decomposition is combined with Tapio decoupling analysis to explore the decoupling relationships between CO2 emissions and the above factors. Finally, the findings related to the impacts of technical and non-technical factors are novel.