Demographic structure factors are the key drivers affecting carbon emissions. Few previous studies related to carbon emissions intensity (CEI) have combined decomposition analysis with considerations of the structural evolution of demographic factors. Analysis of the effect of the structural distribution and evolutionary path of human capital on CEI is essential to formulate targeted cleaner production strategies. A spatial panel lag model and a quantile regression model are established using panel data for 30 provinces in China from 1998 to 2017. The models explore the effects on CEI of primary, knowledgeable, skilled, and institutional human capital and their regional heterogeneity. The results demonstrate the following. (i) Heterogeneous human capitals drive stock, quality, and factor allocation effects on reductions in CEI in their process of evolution. (ii) In the eastern region, the regression coefficients of primary, knowledgeable, skilled, and institutional human capital are all negative, as are those in the central and western regions. (iii) The effects vary in five quantiles of CEI for heterogeneous human capitals. The main effect of human capital on reduction of CEI in the eastern region changes from higher to lower CEI quantiles as human capital evolves from the primary to the institutional level. The effect of the reduction in CEI of human capital in the central region is stronger in the higher quantiles, and decreases with upgrading of human capital. There is little alteration in the coefficients during the evolution of human capital in the western region. In the northeastern region, the positive effects are obvious. To reduce CEI, the strategy of matching and developing heterogeneous human capital in combination with economic elements is advanced.