Buildings are considered as one of the most significant sources of energy use and greenhouse gas emissions. However, few studies have estimated fine-scale energy consumption in the building sector, especially at the city level. This study develops a top-down approach based on statistical and geospatial data to estimate building energy consumption with a high resolution (1 km × 1 km) at the city level. Two representative cities, i.e., Beijing and Shanghai, were chosen to validate the practicality and applicability of the proposed approach. Highly detailed maps of building energy consumption and energy intensity with a resolution of 1 km were generated. The results reflect the spatial non-equilibrium characteristics of building energy use at fine scales. In addition, based on three urban morphology indicators (i.e., building coverage ratio, floor area ratio, and building height), varying relationships between building energy consumption and urban morphology in different cities are revealed. The findings of this study may provide helpful scientific evidence for policy makers to develop appropriate energy-saving strategies for the building sector.