In order to achieve the structural compactness and lightweight design of pole-climbing robots, and overcome the difficulty of vertical pole movement, the weight reduction can be realized by reducing the thickness of parts. However, it can cause local stress concentration, resulting in insufficient stiffness of robot structure. Therefore, starting from the morphology of inchworms, the SIMP (solid isotropic material with penalization) variable density topology optimization design was carried out for the main structure of the pole-climbing robot, realizing the lightweight design of the robot while ensuring its overall performance. Firstly, the main components such as the gripper and T-joint of the pole-climbing robot were selected as the objects, and three motion gaits were dynamically simulated by ADAMS simulation software to simulate the load changes caused by the three motion gaits under extreme working conditions. Then, the boundary conditions were set based on information such as extreme working condition loads. The Topology Optimization module in ANSYS Workbench software was used to perform SIMP variable density topology optimization on the pole-climbing robot to remove the redundant materials that were slightly affected by loads from the robot structure. The optimized robot model was reconstructed and compared with before optimization. The results showed that while the overall weight of the pole-climbing robot was reduced by 7.6% (from 13.60 kg to 12.57 kg), the comprehensive performance of the gripper and T-joint with larger force was improved. The experimental results of energy consumption test showed that the operating energy consumption of the optimized pole-climbing robot decreased by 7.0% compared with before optimization. The proposed SIMP variable density topology optimization method has a high reference value in the structural design of biomimetic pole-climbing robots.