Robotic peg-in-hole tasks are fundamental to many industrial applications but remain challenging due to the need for precise component alignment. Vision based systems enable effective object detection, but their pose estimation accuracy often degrades in the presence of hardware noise and environmental disturbances. To address this, we present Tactile-guided Exploration and Positioning (TacEP), a framework that integrates visual and tactile sensing for high-precision positioning of the target receptacle in robotic peg-in-hole tasks. TacEP first estimates the receptacle’s pose from a monocular camera. To compensate for errors during camera estimation, the system reactively adjusts the pose of a high-resolution Vision-Based Tactile Sensor through real-time exploration, enabling finer positioning of the receptacle. We evaluated TacEP on a physical robotic system across peg-in-hole tasks with five daily connectors, achieving an 81.0% success rate with mean position and orientation errors of 0.37 mm and 0.48◦, respectively. Compared to the vision only baseline, TacEP achieved a 97.7% reduction in position error and an 89.5% reduction in orientation error. These results demonstrate that the proposed framework significantly improves insertion reliability and accuracy, highlighting the potential of tactile sensing for high-precision robotic peg-in-hole tasks.