The integration of Building Information Modelling (BIM) and Geographic Information Systems (GIS) through open standard models such as IFC and CityGML is fundamental to the development of a comprehensive digital twins for the built environment projects. Accurate georeferencing of these models is essential for performing reliable geospatial analyses, particularly in projects that employ a custom Coordinate Reference System (CRS) or are affected by significant distortion due to map projection. Although a considerable number of studies addressed georeferencing in the context of BIM-GIS data integration, limited attention has been given to the assignment of custom CRSs to both IFC and CityGML models. Moreover, the impact of BIM modelling and different georeferencing approaches on the final positional accuracy of the models remains under-investigated. This study addresses these gaps by proposing a methodology to enhance georeferencing accuracy through the assignment of custom CRS to both IFC and CityGML models, while also practically examining how BIM modelling practices and georeferencing approaches impact the overall positional quality of the models. The results demonstrate the effectiveness of the proposed methodology in enhancing georeferencing accuracy of IFC and CityGML models, offering a practical guideline for ensuring the spatial accuracy of the 3D models in the digital twin environment.