This paper examines the carbon reduction effects of environmental regulation (ER) and its dynamic spatial interactions by combining the Stochastic Impacts of Regression on Population, Affluence, and Technology (STIRPAT) model with the dynamic spatial Durbin model (DSDM). The findings indicate that ER can reduce carbon intensity (CI) in China, exhibiting significant spatiotemporal interactions. The reduction is most noticeable in China’s eastern and middle regions but not in the west. Moreover, the mechanism test shows that, in the short term, ER optimizes energy structure (ES) and industrial structure (IS) and promotes technological innovation (Pat). Meanwhile, ES has the best emission reduction effect (-20.54%), followed by Pat (-12.55%) and IS (-2.23%). Only Pat can contribute to long-term pollution reductions. Specifically, the eastern region reduces CI primarily through IS improvements, while the middle region achieves this mainly through ES adjustments. Finally, the paper makes focused policy suggestions for effectively implementing ER to reduce CI and promote high-quality economic development.