Abstract:The effect of spatial spillover and technical spillover between regions is increasing daily, and the study on energy intensity convergence should not overlook spatial dependence and time dynamics. Based on considering the geographic neighborhood of spatial dimension and the dynamics of time-dimension, this paper uses dynamic spatial panel model to analyze the spatial distribution, dynamic change trends and convergence of the energy intensity in 29 provinces and municipalities of China, and the research shows that the spatial distribution of energy intensity of China demonstrates the pattern with low intensity in the east and high intensity in the west and has spatial autocorrelation phenomenon, that the dynamic trends indicates that the spatial autocorrelation index of energy intensity is inclined to be stable after 2005, that the energy intensity in the east, in the middle area and in the west of China demonstrates declining trend, that in overall trend, there is an obviously absoluteσconvergence and no absoluteβconvergence but there is a significant conditionalβconvergence. In order to raise energy efficiency to shrink regional energy efficiency gap, China should sufficiently make use of spatial spillover effect, encourage the technical spillover from the regions with low energy intensity to the areas with high energy intensity so that the convergence speed of energy intensity between regions of China can be accelerated.