Non-destructive disassembly plays a crucial role in remanufacturing, component repurposing, and recovery, which receive great attention due to current circular economy aspirations. As screwing is one of the most common assembly operations, reliable robotic screw unfastening is a key technology in automated handling of end-of-life (EoL) products. Existing approaches, however, show varying reliability, partly due to screw localization difficulties and partly because of precision-related challenges, largely caused by tool tip uncertainty. Addressing the precision aspect, this paper introduces a 2-step visual servo-based method that compensates for tool tip position variability of powered screwdrivers together with uncertainties from imprecise fixturing, robot precision, and manufacturing tolerances. Employing the authors’ patent-pending point-to-line finepositioning technique with two low-cost cameras mounted on the robot, the bit axis of the robot-grasped powered screwdriver is iteratively aligned to the target screw. By performing visual servoing on two different planes subsequently, the geometric relation between robot tool center point (TCP) and bit axis is determined, ensuring reliable bit–screw engagement. The method is presented through an EoL electronic product disassembly case study, achieving 100% success rate, significantly outperforming other potential positioning strategies for robotic screw removal.