Singapore, branded as a “City in a Garden”, has a long standing commitment to green the nation, one which has resulted in trees becoming an integral component of the urban environment. Similarly for its digital twin, Virtual Singapore, we undertake the research to automate the population of this virtual city with semantically and biologically representative trees in a CityGML (City Geography Markup Language) format. This paper presents our framework of modeling trees for Virtual Singapore, showcasing an array of methodologies in data acquisition of light detection and ranging (LiDAR) and satellite images, tree extraction and quantification, and 3D tree modeling at LODs (level of details) 1, 2 and 3. The paper will also highlight challenges and chosen methodologies along with the preliminary results of this framework.