Abstract: In the current development of smart cities, the integration of Building Information Modeling (BIM) and Geographic Information Systems (GIS) faces challenges such as the difficulty in parsing multi-source heterogeneous data and limitations in knowledge representation and reasoning capabilities. In this paper, BIM and GIS data are upgraded to semantic level fusion, and a hierarchical fusion architecture method is proposed to solve the technical problems of geographic information ontology and building information model ontology fusion. This architecture comprises four layers: data, logic, platform, and application. The data layer is responsible for parsing and storing multi-source data; the logic layer constructs ontologies and knowledge graphs for geographic and building information. The platform layer integrates model management with 3D GIS functionalities. The application layer applies the integrated outcomes to fields such as engineering planning and design. By unifying key identifiers and employing various tools collaboratively, this research achieves the integration and visualization of knowledge graphs combining macro-level geographic elements with micro-level building structures. The research results verify the feasibility of the technical path. The proposed method provides a new idea for the construction of macro and micro data fusion and integrated data platform of smart city, and promotes the development of fine management of smart city.