Abstract: The harvesting of Lycium barbarum L. ( L. barbarum) is highly seasonal, with multiple harvesting cycles required per season, and it involves intensive manual labor. Existing L. barbarum harvesting machines face significant limitations: those that achieve high net picking rates typically operate at low efficiency, while machines designed for high efficiency often cause excessive fruit damage. These limitations not only reduce the economic efficiency of L. barbarum cultivation but also restrict the adoption of mechanized harvesting methods in commercial production. To address these challenges, this study proposed a vision-based precise vibration harvesting method and developed a precise multi-point vibration L. barbarum harvester. The proposed precise vibration harvesting method was specifically designed for L. barbarum grown in a double-layer hedgerow cultivation system, which was widely adopted in commercial orchards. The method relied on machine vision technology to detect ripe fruit regions and then clustered these regions into suitable harvesting zones. For each zone, the optimal vibration picking point was determined. By precisely detecting ripe fruit regions, this approach minimized mechanical contact with ripe fruits, thereby reducing fruit damage, while simultaneously ensuring harvesting efficiency. In this way, the method achieved a coordinated optimization of harvesting precision and efficiency, addressing the trade-off that had previously limited the performance of conventional harvesting machines. This paper systematically described the overall structure and working principle of the harvester, highlighting the design and theoretical analysis of key components, including the multi-point vibration picking device, the three-axis harvesting platform, and the cross-row crawler chassis. Based on the growth characteristics of L. barbarum and practical harvesting requirements, critical parameters were determined. These included the layout, length, material, vibration frequency, and torsion angle of multiple vibration ends in the picking device; the working space of the three-axis harvesting platform and the architecture of the control system; and the main structural and driving parameters of the cross-row crawler chassis. Collectively, these parameters ensured effective transmission of vibration energy, stable operation of the platform, and precise interaction between the harvesting unit and the plant canopy. Additionally, a comprehensive control process for precise vibration harvesting was developed, providing a feasible technical solution for intelligent mechanized harvesting of L. barbarum. To optimize harvesting performance, a three-factor, three-level orthogonal experiment was conducted to investigate the effects of excitation position, excitation time, and lateral offset on the harvesting rate of ripe fruits, the harvesting rate of unripe fruits, and the damage rate of ripe fruits. Mathematical models were established to quantify the relationships between these factors and the evaluation indicators, providing a scientific basis for determining the optimal operational parameters. The analysis identified the optimal parameter combination as an excitation position of 55.14 mm, an excitation time of 3 s, and a lateral offset of 30 mm. Performance tests of the harvester demonstrated that a single precise multi-point vibration harvesting unit achieved a ripe fruit harvesting rate of 90.85%, an unripe fruit mis-picking rate of 2.86%, and a ripe fruit damage rate of 5.34%, with an operational efficiency of approximately 19.6 kg/h. By implementing a two-sided, double-layer, four-station layout, the synchronous harvesting efficiency was significantly increased to 78.4 kg/h, which was approximately 10 to 16 times higher than that of manual harvesting. The proposed precise multi-point vibration harvester showed excellent performance in selectively harvesting mature fruits while leaving unripe fruits intact, improving efficiency, reducing fruit damage, and effectively achieving precise and high-efficiency mechanized harvesting of L. barbarum. Overall, this study provided not only a practical and scalable technical solution for L. barbarum mechanization but also a foundation for future intelligent harvesting systems. The combination of vision-based detection and precise vibration harvesting offered valuable insights for improving mechanization, reducing labor intensity, and enhancing both the quality and efficiency of commercial L. barbarum production. The precision multi-point vibration-based L. barbarum harvester developed in this study demonstrated the feasibility of precise and efficient mechanized harvesting of L. barbarum, providing reliable technical and equipment support for intelligent L. barbarum harvesting.