Brain connectome analysis provides insights into body mass index (BMI)-related brain topology and cognitive functions. While alterations in the brain connectome have been observed in individuals with high BMI, evidence regarding BMI-based structural connectome alteration remains limited. In this study, we analyzed diffusion magnetic resonance imaging tractography-derived structural connectivity from 283 neurologically healthy participants by generating low-dimensional features using dimensionality reduction techniques. Two key metrics were calculated: manifold eccentricity, which indicates the relative distance of each brain region from the center of the low-dimensional manifold space, and manifold differentiation, which represents the distance between brain regions within the manifold space. Our findings revealed that individuals with high BMI exhibited greater expansion and differentiation in the control, default mode, and somatomotor networks, reflecting increased network segregation. In contrast, the visual and limbic networks displayed higher integration. Furthermore, network communication measures based on search information and path transitivity indicated less efficient communication between low-level sensory and higher-order transmodal networks in individuals with high BMI. Finally, significant associations were identified between the manifold features in the prefrontal and somatomotor regions and eating behaviors, as assessed by self-report measures from the Eating Disorder Examination Questionnaire (EDEQ) and the Three-Factor Eating Questionnaire (TFEQ). These results highlight the critical role of structural connectome organization in describing BMI and eating behaviors.