Abstract: Global warming and the energy crisis have recently gained growing concern. Urban researchers increasingly focus studies on energy-efficient cities from varying perspectives, especially the urban form's perspective, to gain insight and improve efficiency contributions in overall design components. As micro-level studies taking family or individual as research unit have yielded abundant accomplishments, relatively the much fewer macro-level studies taking cities as research unit have mainly focused on urban scale and density's impact on energy consumption. The relationship between energy consumption and urban spatial structure, such as centrality of facilities, mixing degree of landuse and distribution of block areas, is still lack of proof from empirical research since gathering macro-level urban spatial data is much more difficult than micro-level study. Contributed to the supplement of these empirical proof, this article first makes systemic quantitative expression of urban spatial structure of 30 provincial capital cities in China using the method of Kernel density analyses and landscape index analyses with the data from OpenstreetMap and other online maps, followed by analyses of the relationship between household transportation energy consumption (HTEC) and urban spatial structure. Some findings and suggestions are given: ① From centrality perspective, debate between monocentric form and polycentric form doesn't take up the vital point, a certain center in the urban space hierarchy system should cover proportional built-up area. When an employment center covers about 140 km2 built-up area, a shopping center covers about 45 km2 built-up area and a residential patch covers no more than 1.2 km2, HTEC is the lowest in our study. ② From the perspective of mixing degree of landuse, different from the common sense that the higher mixing degree the better, the average landuse patch area and HTEC are well fitted with a conic curve, meaning that a city whose average landuse patch area is about 5 hectares has the lowest HTEC. ③ From the perspective of distribution of block area, cities with higher proportion of small blocks tend to have lower HTEC.