Orchard tasks often require robots to navigate along the central axes of tree rows in the dynamic orchard environment, making traditional map-based path planning methods unsuitable for these environments. In this work, we propose a practical autonomous navigation solution tailored for orchard operations. We decompose the orchard navigation task into multiple straight-line segments, with the starting and ending points of each segment located on the central axis of the tree row. The waypoints are computed with a trunk map that contains only the trunk point clouds. In the actual navigation process, we sequentially connect all target points to form the robot’s global path and utilize the Dynamic Window Approach (DWA) algorithm for path tracking and obstacle avoidance. Our experiments in real-world environments validate the feasibility of our autonomous navigation method. The proposed navigation approach does not require real-time detection of fruit trees and is capable of dynamic obstacle avoidance, thus offering broad application prospects.