Abstract: The insufficient automation and intelligentization of coal mine roadway support equipment has severely hindered the improvement of roadway formation efficiency, emerging as one of the critical factors contributing to the imbalance of excavation and support. To address the challenges of low automation levels and poor support efficiency in existing roadway support systems, this study has developed an anchor drilling robot integrating a boom-type roadheader with a multi-degree-of-freedom manipulator, proposing a step-by-step visual servo-based control methodology. Firstly, a binocular vision-based spatial positioning method for drilling holes was established utilizing triangulation principles to achieve precise hole localization. Secondly, through displacement and angular sensors deployed on manipulator joints, this study constructed an enhanced kinematic model using the modified Denavit-Hartenberg (MDH) method, enabling accurate spatial position resolution of the end-effector and determination of its initial and target positions. Thirdly, a novel step-by-step visual servo control strategy was developed: a position-based visual control model governs the manipulator's coarse positioning near target holes, followed by an image-based visual servo model that establishes mapping relationships between feature point velocities (derived from circumscribed rectangle vertices of drilling holes) and joint velocities, enabling precise centering alignment. Finally, an experimental platform was established to validate the proposed methods under laboratory conditions. Test results demonstrate average positioning errors of 7.5, 7.8, and 8.1 mm in X, Y, and Z axes respectively, with the step-by-step control approach successfully achieving both coarse positioning and fine alignment. This research significantly enhances the automation level of support systems, laying a technical foundation for workforce reduction and efficiency improvement in roadway support operations.