Building information modeling (BIM) is a new technology that supports lifecycle information management for buildings. Ontology, derived from philosophy, is another technology utilized to describe real-world entities and to infer their semantic relationships. Each year disasters such as fires, earthquakes and floods affect not only buildings but all occupants inside, and hence, identification of critical building elements and devices vulnerable to a disaster is a must in order to perform countermeasures to ensure the safety of lives and property. At present, many simulation-based tools such as FDS have been successfully employed to mimic the impact of a disaster. Nevertheless, most of such approaches consider only the spatial or geometry data of a building. Other building-related attributes such as materials and equipment that certainly influence how a disaster proceeds have not been thoroughly examined. Thus, this research develops BIM plug-in programs to extract not only geometry but building materials and equipment/devices information in order to craft a disaster-specific ontology. Semantic relationships are modeled using Semantic Web Rule Language (SWRL) constructs, so as to represent the interdependencies among building elements and devices under a given disaster. The combined approach can accommodate not only spatial relationships but functional constraints pertaining to internal building elements and devices. It may help disaster mitigation officials forecast and assess the impact and progress of a disaster. A train station was utilized to demonstrate the proposed approach, with focuses on the fire and flood types of disasters. Better disaster mitigation strategies can be prepared if the proposed approach is adequately utilized.