Ecological restoration of degraded areas is commonly an expensive enterprise that can result in varying levels of recovery. With limited financial resources, the restoration efforts should focus on areas where restoration will produce the greatest benefits. There are many coal subsidence areas in the coal cities on the eastern plain in China, and these areas have harmed the structure and functioning of urban green infrastructure (GI) for a long time. Setting restoration priorities is also an urgent task for restoration practitioners and urban planners in coal cities, and identifying the restoration objective is the key for assessing restoration priorities. Many studies on coal subsidence areas focus on local objectives to evaluate the land suitability (e.g. soil restoration, habitat restoration). However, few studies so far have considered a broader context, such as the perspective of sustaining GI stability, when planning restoration actions. In this paper, we investigate which coal subsidence areas should have priorities for restoration, based on the goal of GI. We consequently propose a framework for prioritizing the restoration actions of coal subsidence patches, based on the ecological suitability of GI and the potential for sustaining urban landscape connectivity. Coal subsidence patches with high ecological suitability and with enough potential capacity to favor urban landscape connectivity were given a higher priority, and could therefore play a strong positive role in improving urban GI stability and sustainability. Our developed method consisted of three steps. First, we quantified the ecological suitability within each coal subsidence patch, using a vertical overlay method. This method takes into account variables such as land use status, normalized difference vegetation index, distance to the nearest ecological protection area, distribution of the main rivers in the patches, and distance to grey infrastructure and industrial areas. Second, we identified the spatial relationships between the coal subsidence patches and existing ecological patches, and subsequently quantified their potential capacity for improving the urban landscape connectivity using Conefor Sensinode 2.6 software. Finally, we overlaid the results of the ecological suitability and the potential for sustaining connectivity, giving both scores equal weights. The coal subsidence patches were then ranked according to their total score. To demonstrate an application of our proposed method, we present a case study for Xuzhou, Jiangsu Province. More than half of the coal subsidence areas here demonstrated a high ecological suitability for GI, and these areas were distributed in the region with a good ecological status and little human disturbance. The coal subsidence patches in North and Central Jiawang had significant roles in sustaining landscape connectivity. Finally, the coal subsidence area restoration priority values were divided into five grades, with the high and higher priority sites generally located in Central Jiawang and Central Pangzhuangdong, and in North Dongzhuang and North Zhangji. Overall, our proposed method demonstrated to be feasible when used for prioritizing restoration actions of coal subsidence areas in Xuzhou.