Autonomous ground vehicles have the potential to reduce the risk to Soldiers in unfamiliar, unstructured environments. Unmanned operations in unstructured environments require the ability to guide the vehicles from their starting position to a target position. This paper proposes a framework to plan paths across such unstructured environments using a priori information about the environment as cost criteria into a multi-criteria, multi-agent path planner. The proposed multi-criteria, multi-agent path planner uses a penalty-based A* algorithm to plan multiple paths across the unstructured environment and uses entropy weighting for generating weights to calculate a multi-criteria cost with distance, risk, and soil trafficability. The paths generated by the proposed framework provide a better overall performance across the cost criteria and can be used as waypoints to navigate UGVs in off-road environments.