More recently, the COVID-19 pandemic outbreak has created massive economic policy uncertainty (EPU). EPU and its economic fallout have been a hot topic of study, however, the impact of EPU on CO2 emissions has been seldom addressed to date. This paper investigates the direct impact of the EPU on CO2 emissions and indirect effect via the environmental regulation at the national and regional levels using the panel data model and provincial panel data from 2003 to 2017 in China. The empirical results show that the central region is the most special one, which all explanatory variables except energy consumption are all non-significant even at the 10% level. For other samples, there is a significant positive correlation between EPU and CO2 emissions, whether in the national or regional level. Additionally, environmental regulation alone can achieve the purpose of curtailing carbon emissions. However, when the EPU is taken into consideration, environmental regulation exerts a significantly positive effect on CO2 emissions, leading to unintended increase in emissions. Moreover, the Environmental Kuznets Curve (EKC) hypothesis was confirmed in the national and eastern samples, while CO2 emissions increase monotonically as economic level grows for western datasets. Based on the overall findings, some policy implications were put forward.