Abstract: The inter-row passages in natural rubber plantations are narrow, with densely distributed obstacles and limited path safety margins, which impose high requirements on the navigation capability of rubber-tapping robots. To address this problem, an improved A* path planning method is proposed. First, the Chebyshev distance is adopted as the heuristic function to optimize the node search process. An obstacle-repulsion-based post-processing mechanism is then introduced to refine the initially generated path for safety distance correction and smoothing. Subsequently, a cooperative navigation framework integrating global path planning and local dynamic obstacle avoidance is constructed, enabling stable navigation and fixed-point stopping of the robot in rubber plantation environments. Simulation results show that the improved algorithm generates paths with fewer steps while maintaining a larger safety distance from obstacles. In the localization accuracy experiment, the LIO-SAM algorithm is quantitatively evaluated using Absolute Pose Error (APE) and Relative Pose Error (RPE). The root mean square errors in the natural rubber plantation environment are 24.80 cm and 22.73 cm, respectively, indicating stable localization performance without noticeable drift. Field navigation experiments further demonstrate that the root mean square errors of both longitudinal and lateral positioning errors are less than 11.00 cm, and the average stopping error is approximately 5.00~6.00 cm.