Rural road planning in mountainous areas must coordinate existing-road use, hierarchical connectivity, and construction requirements. This study proposes a network generation method integrating existing-road topology with roadless space through a hexagonal occupancy grid. Existing-road neighborhood and cross-neighborhood edges represent different access characteristics. Sequential rules generate a county-level backbone, township-level links, and village-level access, while identifying reuse, upgrading, and new construction. Six scenarios are compared in Zhangjiachuan Hui Autonomous County, Gansu, China. The proposed scenario generates an 880.001 km network, with existing roads accounting for 94.53%. New construction totals 48.128 km, 85.63% less than under uniform cross-neighborhood representation. Mean shortest travel times for village-to-township government, township government-to-county seat, and village-to-county seat connections decrease by 15.65%, 37.85%, and 27.91%, respectively, relative to the pre-planning network. All three hierarchical compliance rates reach 100%. The proposed scenario achieves the highest composite score among the six scenarios under the specified indicators and weights, balancing existing-road use, service accessibility, and construction requirements. The method supports preliminary network layouts and scenario comparison in mountainous counties with comparable terrain and road data.

Sustainable rural road planning in mountainous areas must coordinate existing-road use, hierarchical connectivity, and construction requirements. This study proposes a network generation method integrating existing-road topology with roadless space through a hexagonal occupancy grid. Existing-road neighborhood and cross-neighborhood edges represent different access characteristics. Sequential rules generate a county-level backbone, township-level links, and village-level access while identifying reuse, upgrading, and new construction. Six scenarios are compared in Zhangjiachuan Hui Autonomous County, Gansu, China. The proposed scenario generates an 880.001 km network, with existing roads accounting for 94.53%. New construction totals 48.128 km, 85.63% less than in E1, which represents all existing roads as cross-neighborhood edges and reuses them directly. On the full post-planning evaluation network, mean shortest travel times for village-to-township government, township government-to-county seat, and village-to-county seat connections decrease by 29.69%, 45.24%, and 37.06%, respectively, relative to the existing-road network. All three hierarchical compliance rates reach 100%. Using entropy weights of 11.90%, 44.82%, and 43.28% for network composition, service accessibility, and construction requirements, respectively, the proposed scenario achieves the highest composite score among the six scenarios. The method supports resource-efficient rural transport planning by coordinating existing-road use, hierarchical service connections, and new-construction requirements in preliminary layouts for mountainous counties with comparable terrain and road data.