<p>Under the strategic goal of carbon peaking and neutrality, accelerating industrial carbon decoupling is one of the key duties for achieving high-quality economic development. This paper focuses on the decoupling effect and decomposition of carbon emissions from the standpoint of energy structure optimization. Based on data from 17 energy categories in 26 industrial sub-sectors at the provincial level in China from 2005 to 2021, we calculate the industrial carbon decoupling index. Then, an integrated method “Tapio + LMDI” is constructed to analyze the decoupling status and its driving factors. The study finds that: (1) The industrial economic growth and carbon emissions in China are in a weak decoupling state. Specifically, most provinces have achieved either weak decoupling or strong decoupling of carbon emissions. A few provinces are in a disadvantageous position of decoupling, mainly concentrated in the northern regions. (2) From the decomposition results, it can be seen that the economic output has become the main factor driving the increase in industrial carbon decoupling. Conversely, the energy structure effect can decrease carbon decoupling, but its contribution is relatively weak, accounting for only 16%. (3) At the regional level, energy structure plays a “supplementary role” in most provinces. The contribution of energy structure optimization to carbon decoupling is usually lower than that of industrial structure optimization or technological progress. (4) At the industry level, energy structure optimization is the conventional way for most energy-intensive and high-emission industries to achieve carbon decoupling. These findings not only provide new empirical evidence for the intrinsic relationship between energy structure and carbon decoupling, but also offer feasible insights for industrial carbon emission reduction in developing countries.</p>