High-density areas are facing the problems of a lack of eco-space and fragmented layouts due to rapid urbanization. This study, in analyzing the main issues in the planning domain, proposes that roof greening as a complementation can be integrated into urban green systems, and aims to construct an eco-space network in a high-density city area with a deteriorated ecological environment and shortage of land resources. The present study uses Xiamen Island as the research area, and utilizes geographic information system (GIS) and remote sensing technology to establish the status of the area and plan the space database by using remote sensing data, field surveys, questionnaires, departmental data, and planning texts. From the three aspects of existing research combing, comprehensive benefits, and urban development appeal analysis, this study summarized the impact factors of roof greening implementation potential, integrated planning divisions, and building classifications to establish a method of assessing implementation potential; we assessed the potential of roof greening implementation in the study area, and achieved eco-space network planning combined with urban green-space systems by taking "global space, key areas, and multidimensional networks" into consideration. The results indicated that 1) ecological structure, urban function optimization, public service demand, environmental improvement demand, architectural characteristics, and architectural property ownership as the evaluation elements of potential for roof greening implementation, including the urban ecological framework, urban functional structure, landscape upgrade demand, building renovation demand, urban internality degree, heat island intensity, microclimate environment, building height, building age, building type, and architectural property attributes as key evaluation factors. 2) In 2016, the roof greening rate of buildings on Xiamen Island was only 2%, with more than half attached to residential buildings and urban villages. Commercial, public, and industrial buildings had a lot of potential space for roof greening, and among the remaining buildings, 81.3% were suitable for roof greening. 3) The first step of the study was to explore the method of planning the eco-space network in the high-density urban area. Next, with the help of artificial intelligence (AI) and GIS, we aimed to build a research method for optimizing an eco-space network with multiple functions, in order to prevent the urban environmental damage and ecological disaster caused by high-intensity development and high-risk agglomeration.