Abstract: The primary use of robotics in nuclear decommissioning applications is to reduce the radioactive dose levels to which workers are exposed during the decommissioning processes. There are many situations where, owing to the degree of radiation and the very long half-lives of the radioactive materials involved, robotics is the only feasible option. However, unstructured decommissioning environments pose significant challenges to the navigation, operation and control of tele-operated mobile robots and more specifically multi-arm mobile robots. Through task planning is often not achievable due to the risk of radiation and the unpredictable nature of the decommissioning workspace environment. The hazardous and unpredictability of the decommissioning environment makes real-time sensing with vision system and decision making an essential component of the multi-arm mobile robot. This paper proposes the use of a sensing strategy and visual feedback from cameras for navigation and position estimation and to provide the operator with real-time information of the target and its surrounding for controlling the multi-arm mobile robot and the successful execution of the decommissioning task.