The working environment of wheeled robots has gradually evolved from indoor to outdoor. In order to meet the requirements of safe and low-energy autonomous task execution by wheeled robots in outdoor uneven terrain environments, this paper proposes a multi-layer 2.5D grid map, which enables robots to avoid obstacles while avoiding undulating terrain that they cannot cross. To reduce the energy consumption of robots performing tasks outdoors, this article first analyzes the impact of a single path on the energy consumption of the robot, establishes a robot path energy consumption model, and introduces this path energy consumption model into the cost function of the A* algorithm to obtain the PE-A * algorithm. The PE-A * algorithm is used to plan a single low energy consumption path for the robot; Secondly, when a robot has multiple task points, this article abstracts the robot's access to multiple task points as a TSP problem to find the access sequence that minimizes the robot's energy consumption; Finally, the feasibility of the proposed method in outdoor environment adaptability and energy consumption reduction has been demonstrated through experiments.