This paper constructs an optimization model for selecting on-site haul roads during rockfill dam construction based on a strategy of interval-discretized elevation adjustment

to address the issue of suboptimal haul road routing

which is frequently exacerbated by steep terrain and slope constraints. This model

rooted in the theoretical framework of Least-Cost Path Analysis (LCPA)

incorporates critical factors－such as path length

elevation modification

and turning constraints－while leveraging the Dijkstra's algorithm for its efficient calculations of a solution. A comparative case study reveals that in comparison to conventional methods

it reduces the haul road length by 17.33% and the overall cost by 16.17%

improving significantly the cost-effectiveness and rationality of the routing scheme

and thereby validating its performance. And its applicability is further exemplified through a case study of the Cihaxia high concrete-face rockfill dam on the upper Yellow River mainstream. The method obtains a route of construction that agrees closely with the engineer-designed path that was refined using multiple iterations

and thus it would be an innovative solution to the planning of haul road in complex terrains.