Abstract: Background Optimizing the spatial layout of vegetation is an important measure to enhance the effect of soil and water conservation. Methods Taking the Lvergou Watershed as the research object, six typical vegetation configuration scenarios (based on elevation and slope gradient) with 40% coverage were set up, combined with SWAT model and improved sediment connectivity index (ICZQ) to analyze the sediment reduction effect of vegetation pattern. Results 1) The SWAT model demonstrated satisfactory accuracy in simulating monthly runoff and sediment transport, confirming its suitability for scenario-based sediment analysis. 2) Vegetation configurations concentrated in higher-altitude areas or gentler slopes exhibited superior sediment reduction, with effectiveness significantly influenced by rainfall conditions. 3) Altitude-based configurations formed continuous vegetation belts, creating large-scale low-ICZQ zones, while slope-based configurations resulted in fragmented patches with scattered low-ICZQ areas. Annual sediment transport showed a significant negative exponential correlation with mean ICZQ across all scenarios (P<0.01). 4) At the sub-watershed scale, a highly significant linear relationship (P<0.001) was observed between changes in ICZQ (ΔICZQ) and sediment yield modulus (ΔSM). Conclusion Optimizing vegetation spatial layouts by regulating sediment connectivity can effectively enhance erosion control, providing a scientific basis for vegetation management in soil and water conservation practices across the Loess Plateau.