Abstract: Large-scale cabin segment docking assembly is a critical process in the integration of major equipment for aerospace, marine, and related industries. Traditional methods depend on manual visual inspection of docking interfaces and human-adjusted segment alignment, which suffer from inefficiency, high labor intensity, and inconsistent quality. To address these limitations, this study proposes an intelligent vision-guided docking system. Utilizing a monocular camera to dynamically capture target markers on cabin surfaces, the system calculates relative 6-DOF (six-degree-of-freedom) poses via spatial geometry algorithms, enabling actuator-driven autonomous docking with high precision. This approach offers distinct advantages, including non-contact measurement, algorithmic simplicity, and ease of maintenance. The paper details applications of this monocular vision-based technology across three representative product models under varying docking scenarios. Experimental results demonstrate a docking accuracy of ±0.37 mm, significantly advancing the automation level of large cabin segment assembly.