Radiation risk represents a major safety concern during the decommissioning of nuclear facilities, necessitating that nuclear decommissioning robots effectively minimize radiation exposure. Path planning plays a critical role in radiation protection by enabling robots to select optimal routes that avoid radiation sources as much as possible, thereby reducing radiation dose accumulation. In this study, a dynamic constraint–based dynamic constraint A (DC-A) star algorithm is proposed to address the path planning problem for nuclear decommissioning robots.

The algorithm explicitly considers the actual boundary dimensions and rotation angles of the robots, optimizing the planning process to ensure task completion with enhanced safety and efficiency. The algorithm enables smooth narrow passage traversal through two coordinated functions: real-time feasibility evaluation of inter-obstacle regions and dynamic adjustment of the robot’s facing orientation based on these assessments.

To validate the effectiveness of the proposed method, a series of radiation-focused experimental scenarios were conducted. The results demonstrate that the DC-A star algorithm significantly reduces radiation dose exposure while improving task performance and success rates in complex environments, thus enhancing the operational safety of nuclear decommissioning robots.