Abstract: This study aims to improve the autonomous navigation accuracy of the mobile chassis in greenhouse ridge cultivation. A combined control system was developed to navigate the leafy vegetable ridges, particularly with poor straightness. The system consisted of a vision, UWB, and controller module. Firstly, a depth camera was used to capture the images of the leafy vegetable ridge surface in front of the robot. The YOLOv8s-seg algorithm was used to segment the ridge surface of leafy vegetables. The Canny algorithm was used to identify the ridge surface edges. Finally, the least squares method was used to extract the visual navigation lines; A control segmentation line between UWB and visual navigation was constructed for the autonomous navigation and tracking of mobile robots in a monopolistic environment. A mobile navigation strategy was proposed for the crop protection robots using UWB/visual combination. A navigation tracking model was established using a pure tracking algorithm; Finally, autonomous navigation experiments were conducted in actual field scenariosThe experimental results show that when the plant protection machine travels at a speed of 0.4-0.5 m/s, the maximum tracking deviations of visual navigation and UWB navigation are 6.3 and 5.8 cm respectively, and the average errors are 4.7 and 4.3 cm respectively. This verifies the feasibility of the navigation line extraction algorithm proposed in this paper and the reliability of UWB navigation, providing a basis for the combination of the two. The regulation dividing line of the integrated navigation is 15 cm. In the two groups of integrated navigation experiments, the maximum lateral deviations are 15.2 and 15.5 cm respectively, the average errors are 5.9 and 6.1 cm respectively, and the maximum heading deviations are 5.9° and 7.8° respectively. The experimental results demonstrated that the UWB navigation control played a significant positive regulatory role in extreme situations. Navigation deviation and deviation fluctuation were significantly reduced in the integrated navigation. Therefore, the combined navigation improved the navigation accuracy and robustness. The finding can provide technical support to the production efficiency in precision agriculture.