Assessing the urban-scale building photovoltaic (PV) potential is important for designing urban environments, retrofitting existing structures, or integrating PVs with grids. However, few studies have considered high -temporal-resolution simulations, the facade PV potential, and a comprehensive PV model simultaneously; thus, the overall accuracy of the estimation of PV potential may be limited. Therefore, this study developed an integrated framework to assess the urban-scale PV potential of rooftops and facades at high spatiotemporal resolution. The proposed approach integrates an anisotropic sky diffuse model, a vector-based shading calcula-tion method, and a temperature-related PV performance model. The annual PV potential and spatial/temporal characteristics were analyzed in a case study of over 170,000 buildings in Beijing. The results showed that the estimated rooftop PV power generation was 7.55 TWh/y, whereas the facade PV power generation was 18.07 TWh/y, which was 239% of the rooftop PV yield. The integrated model estimated PV yield with higher accuracy than the simplified models by depicting more details. The proposed approach can be applied to the large-scale assessment of future energy systems with increasing penetration of PVs, and the results can support effective policies for the integration of PVs into the built environment in dense cities.