In order to improve the unidirectional linear resource metabolism pattern within communities and optimize the urban resource circulation system, a green productive regeneration method for urban communities based on image 3D reconstruction and multi-criteria decision making was proposed with the help of drone low-altitude photogrammetry technology and using Agisoft PhotoScan, Cloud Compare, and Yaahp. By constructing a 3D point cloud model of the community's built environment and a spatial production potential evaluation model, the suitability of different spaces for production within the community was analyzed, and regeneration design was carried out. The results show that the regenerated community using this method can meet 43.8% of its vegetable supply demand, process 87.28% of its organic waste, supplement 23.46% of its irrigation water needs, and effectively reduce food mileage carbon emissions and building energy consumption. It is concluded that the community's green productive regeneration can significantly reduce its external resource dependency, thus transforming its resource metabolism pattern.