<p>Under conditions of constrained urban renewal resources, identifying priority communities has become a critical challenge for urban governance. However, existing studies on renewal potential remain largely problem or material-oriented and provide limited insight into the heterogeneous renewal trajectories across communities. This study develops a community-level spatial framework that integrates renewal demand and driving forces to assess urban renewal potential. Using multi-source spatial data, the framework combines deep learning, network analysis, the CRITIC weighting method, and K-means clustering to capture both the intensity and structural heterogeneity of renewal potential. A case study of 3,967 communities in central Wuhan shows that communities are predominantly characterized by medium renewal potential (46%), followed by high (33%) and low (22%) potential, with significant spatial variation. Five distinct community types are identified: service-saturated aging inner-city communities, structurally deprived inner-city communities, commercially vibrant mixed-use inner-city communities, newly developed low-activity residential communities, and moderately serviced livable communities. Further analysis shows that overall renewal potential is primarily driven by renewal demand, whereas renewal drivers play a crucial role in shaping divergent development trajectories among communities with similar potential levels. Comparison with officially designated urban renewal units reveals a high degree of spatial correspondence, supporting the practical validity of the proposed framework. The framework provides a scalable analytical tool for identifying priority communities and informing differentiated renewal strategies. By leveraging multi-source urban data, it also demonstrates strong transferability and applicability across diverse urban contexts.</p>