English：

Building energy conservation is the key link to achieve the goal of carbon peak and carbonneutralization, and the importance of regional scale building energy consumption research is becomingmore and more prominent with the development of urbanization. To this end, a machine learning-basedregional-scale green building energy consumption prediction framework is constructed. The conversionfrom IFC to CityGML is realized by FME data conversion technology, and BIM and GIS are integrated.The rich semantic information of BIM is used to make up for its shortcomings in regional scale spatio-temporal analysis, and the classification algorithm is applied to classify the building types to improve theprediction accuracy. Through example verification, the three machine learning regression modelsevaluated can accurately predict building energy consumption, and the speed is significantly better thanthe traditional energy consumption simulation method. This study provides a practical idea for theprediction of energy consumption of green buildings at the regional scale,which is helpful to promote thesustainable development of green buildings.