<p>Evaluating the environmental efficiency of the power industry helps to gain a more accurate understanding of its environmental impacts and efficiency performance. This can provide scientific justification for governments and enterprises to formulate more effective energy conservation and emission reduction policies. Considering the influence of external environmental factors and stochastic factors, this paper proposes an improved three-stage slack-based measure with super-efficiency data envelopment analysis (SBM-SE-DEA) model to evaluate the environmental efficiency of the power generation industry in China’s 30 provincial regions during 2015–2021. The model integrates three-stage DEA model, SBM-DEA model, and SE-DEA model while accounting for undesirable outputs such as carbon emissions and air pollutants. The main findings reveal that regions with a high proportion of renewable energy generation exhibit excellent environmental efficiency, although initial assessments are often overestimated due to external environmental factors. Meanwhile, there are significant differences in environmental efficiency between regions, with the Southwest region at the top of the list due to the widespread use of renewable energy, while the Northeast region is at the bottom of the list due to its high reliance on conventional energy sources. After an in-depth analysis that excludes external environmental and stochastic factors, the environmental efficiency of each region shows a more realistic and stable trend, which provides a solid basis for policymaking. To effectively address climate change and environmental pollution, policy-makers should promote renewable energy, enhance regional cooperation in electricity distribution, and invest in technology and human capital to improve environmental efficiency across provinces.</p>