Green innovation is a critical support to combat climate change arising from greenhouse gas emissions generated by energy consumption. It is an essential way to achieve resource storage, carbon emissions reduction, and sustainable development goals in China. Based on an environmental framework defined as the Stochastic Impacts by Regression on Population, Affluence, and Technology (STIRPAT) model, this study aimed to empirically check the impact of green innovation (GI), per capita GDP (PGDP), population density (PD), environmental regulations (ER), energy consumption (EC), and industrial structure upgrading (ISU) on CO2 emissions (CO2e). For this purpose, a sample dataset covering the 30 provincial regions in mainland China from 2005 to 2019 was analyzed using the Fixed Effects and System Generalized Method of Moment (SYS-GMM) Methodology. The empirical results showed that CO2e in the current period were further aggravated due to the agglomeration effect of CO2e from the previous period. The data analysis indicated that GI, ER, and ISU all exert a significant inhibitory effect on CO2e, whereas PGDP, PD, and EC had a positive effect on carbon emissions when dynamic relationships were analyzed. In the regional heterogeneity test, the current model also revealed that the impact of GI on diminishing CO2e was more pronounced in the east-central region, but not in the west. It is suggested that policymakers in China not only design differentiated policies in response to regional heterogeneity, but also focus on the decisive role of green technology application, environmental protection, and green transformation of industrial structure in curbing CO2e.