Maintenance tasks in distribution networks are often accompanied by hazards associated with high altitudes and high voltages. By utilising robots instead of human operators to perform these tasks, potential risks can be avoided, while productivity can be increased. This research proposes an intelligent power distribution live‐line operation robot (PDLOR) system based on a stereo camera to replace human to complete work. The PDLOR system consists of several key components, including dual manipulators, wireless tools, a visual perception system, an insulated bucket truck, and a ground control terminal. Once the task is confirmed, the real‐time vision system identification and positioning enable the adjustment of the insulated bucket to position the robot correctly for its intended work. The stereo camera plays a crucial role in accurately recognising and estimating the object's orientation. Additionally, a simplified reconstruction is performed within a virtual simulation environment, which aids in collision detection during path planning. After obtaining the optimal path, it is then communicated to the real manipulator for execution. To validate the feasibility of the PDLOR system, field experiments were conducted in actual distribution network scenarios. The results demonstrate that the PDLOR effectively completes single‐phase power‐line connection tasks within a remarkable 10‐min timeframe.