Supporting the semantics of 3-D objects and their spatial relations in database systems has been little addressed in the literature. Despite its importance, most 3-D graphic systems lack this capability, mainly focusing on the visualization aspects of 3-D images. We have developed a 3-D graphic data model based on XML that accommodates the semantics of 3-D scenes. This model offers content-based retrievals of scenes containing a particular object or those satisfying certain spatial constraints on them. The model represents scenes as compositions of 3-D graphic objects with associated spatial relations. Complex 3-D objects are modeled using a set of primitive 3-D objects rather than the lines and polygons that are found in traditional graphic systems. This paper presents the data model and its implementation called 3DGML, an XML vocabulary that we developed for modeling 3-D graphic data. This paper also describes a Web-based prototype database system that we developed to support the data model.