Using unmanned aerial vehicle(UAV)images acquired after the Jiuzhaigou M_S7.0 earthquake on August 8, 2017, and the field damage survey data, characteristics of buildings with different seismic damage levels were analyzed and remote sensing interpretation features of buildings in different damage levels based on UAV images were built. Zhangzha and Heye villages were selected as study area, and the seismic damage levels of buildings were extracted by visual interpretation through UAV images. Then, the remote sensing seismic damage index was quantitatively assessed and compared with ground truth by field survey. The result shows that the remote sensing interpretation of building damage and actual damage degree is consistent. It also shows that the very high spatial resolution UAV images in the disaster area acquired immediately after an earthquake can accurately identify the damage of buildings and provide an important reference for earthquake disaster assessment and emergency rescue decision.