This article uses the Dagum Gini coefficient and kernel density estimation to analyze the regional differences and distribution dynamics of energy consumption and applies the coefficient of variation with a static panel model to test its σ-convergence and β-convergence characteristics. The study finds noticeable differences in energy consumption intensity among provinces and regions, and the main source of regional differences is inter-regional differences. In contrast, the contribution of intra-regional differences and hypervariable intensity is smaller and more similar. Specifically, the absolute difference in the eastern region experienced a first decline, then a rise, and finally a slight decline; the absolute difference in the central region had a more significant overall fluctuation and a declining trend in the later period, and the absolute difference in the western region tended to narrow. No polarization trend was observed during the sample period in the eastern, central, and western regions. Not only σ-convergence but also absolute β-convergence and conditional β-convergence exist in the whole country and the eastern, central, and western regions. Therefore, it provides a factual basis for deepening the perception of the spatial pattern and regional disparity of energy utilization levels in China and has policy implications for promoting inter-regional and intra-regional coordinated development.