This paper examines the endogenous evolution of household consumption patterns and household carbon emissions (HCEs) by integrating the analysis methods of income distribution with climate change. Based on a largescale household survey spanning from 2012 to 2016 in China, we estimated the direct and indirect HCEs, observed inverse U-shaped Carbon Kuznets Curves (CKC) and significant changes in HCEs over the period at the household level. Applying the Oaxaca-Blinder method, we decomposed factors causing the changes in HCEs and found that income and demographic effects contribute only 25.1% to the total increase of HCEs. The other 74.9% remain unexplained and we define them as the effect of intertemporal lifestyle changes. Further analysis from multiple perspectives illustrates that the lifestyles of households across various social strata are becoming increasingly higher carbon-intensive over time even though the income remains unchanged. The findings indicate that existing modeling and projections of carbon emissions based on income and household characteristics may underestimate the future emissions pressure from the household sector. Hence, we conclude that in order to reach more meaningful results, the increasing effect of lifestyles should be taken into account when conducting climate change studies and formulating climate policies.