Coastal erosion vulnerability assessments are critical for planning effective coastal defense; however, previous methods that rely on low-resolution data often lack the analytical precision required for local- or site-specific coastal management. This study presents an improved approach that constructs a 3D geospatial digital twin of coastal areas by integrating high-resolution aerial image data and topo-bathymetric light detection and ranging (Lidar) data, which are acquired by airborne sensors. Using this digital twin, coastal erosion vulnerability was assessed at 5 m intervals along the coastline in the digital twin based on six vulnerability indices including geomorphological feature, beach width, coastal slope, and so on. A key innovation of the proposed method is the application of the Jenks natural breaks optimization to measure vulnerability scores for each index. Then, the vulnerability assessment is determined by combining these six scores. The method was applied to a coastal area in Uljin-gun, Republic of Korea, and, unlike previous low-resolution 2D data assessment approaches, it successfully identified specific coastline segments exhibiting high vulnerability that would not be identified with previous approaches.