<p>Global food security is increasingly threatened by cropland loss and rising food demand, necessitating policy interventions such as settlement consolidation. However, scaling up local consolidation practices requires large-scale assessments of consolidation potential to effectively coordinate interventions for localized implementation. Here, we evaluate settlement consolidation potential across rural land systems in Zhejiang Province, by leveraging machine learning models fed with reference data from completed consolidation projects and associated explanatory variables. This consolidation potential assessment demonstrates high robustness and precision, with a recall of 0.75, a modified F1 score of 4.05, and a low coefficient of variation. The consolidation potential map illustrates heterogeneity of settlement consolidation potential across space, which identifies 29,784 hectares of potential consolidation area in rural land systems, predominantly concentrated in northern prefectures such as Jiaxing (12,961 hectares) and Huzhou (8987 hectares). In general, smaller settlement patches exhibit higher consolidation potential. In addition, a distinct distribution of consolidation potential is observed across plains, hills, and mountains. Specifically, plains account for almost 80% potential consolidation area, followed by mountains (17%) and hills (3%). This study reveals the spatial heterogeneity of settlement consolidation potential, offering insights for strategies that enhance food production, optimize land use, and inform policy-making.</p>