The consumption of electricity accounts for a higher and higher proportion of the terminal energy structure. The efficiency of electricity-use thus should be of great concern, especially under the background of energy saving and emission reduction. Instead of measuring the efficiency of electricity-use via electricity intensity, the input-oriented epsilon-based measure (EBM) model under variable returns to scale, supported by a total-factor framework, is employed to evaluate the efficiency of electricity-use of China's 30 regions spanning from 2006 to 2015. Then kernel density estimation and Moran's I index are used to depict its dynamic evolution characteristics and spatial agglomeration characteristics. Finally, a penalized panel quantile regression model with fixed effects, taking unobserved individual heterogeneity and distributional heterogeneity into account, is applied to estimate the effects of probable determinants on the efficiency of electricity-use. It found that there are significant differences in the efficiency of electricity-use in terms of time and space dimensions, whose distribution seems to be a projection of the level of regional economic development. The regional efficiency of electricity-use exhibits a descending trend and is positively spatially autocorrelated. The impacts of various factors, namely, income level, population size, industrial structure, urbanization and FDI intensity, are obvious heterogeneous throughout the distribution. (C) 2018 Elsevier Ltd. All rights reserved.