A new method applying close range laser scan (Lidar) survey in capturing 3D objects was presented in this study. One linear laser source Lidar was used for the survey of the objects. This approach is particularly essential for 3D modelling of urban buildings in a geospatial database. Data calibration and merging were conducted for combing point cluster datasets surveyed from different locations and angles of one object in order to generate a 3D data model for the entire object. The point cluster data were transformed and converted into a triangulated irregular network (TIN) mesh model in a 3D geospatial database. The TIN mesh model was compressed and simplified by applying a computing script of surface variability priority parameters. The digital surface images and textures of the building object were draped onto the 3D TIN model according to their coordinate locations. This research provides a direct and fast means of capturing 3D objects on the Earth's surface in a geospatial database.