Abstract: Efficient path planning for manipulators is essential in modern CNC machining processes to ensure obstacle-free operations, particularly during complex tasks such as picking, handling, and clamping. The presence of irregularly shaped obstacles at various stages of machining presents significant challenges for traditional path planning algorithms, which often struggle to balance computational efficiency with path quality. To address these challenges, this study proposes a novel, efficient, and high-quality path planning method for loading and unloading manipulators, leveraging the advanced capabilities of digital twin technology. This approach not only improves operational efficiency but also ensures safety and adaptability in complex working environments. Central to this method is the proposed RRT*-Connect algorithm, which performs dynamic elliptical constraint sampling. By integrating dynamic constraint sampling with adaptive step-size adjustments, the algorithm significantly enhances the efficiency of random tree search process. Targeting feasible regions through elliptical constraints accelerates convergence toward optimal paths, while techniques such as redundant node elimination and initial path smoothing further improve path quality, resulting in shorter, smoother routes. To validate this approach, a real-time digital twin simulation environment was constructed. This high-fidelity environment accurately reflects manipulator operations by incorporating real-time data collection and bidirectional transmission of operational parameters. Through the digital twin framework, manipulator movements and interactions with obstacles are precisely mapped, enabling real-time monitoring and waypoint updates. The simulation environment also supports real-time updates of manipulator status, ensuring that the planned paths remain both feasible and optimized for actual operations. The effectiveness of the proposed method was demonstrated through a comprehensive case study on a CNC machining production line. Comparative experiments with two baseline algorithms, RRT*-Connect and an improved RRT*-FN algorithm, highlight the superior performance of the proposed method. Specifically, the proposed approach reduced operation time by 30.68% and 23.56% and terminal path cost by 24.76% and 14.99% compared to the respective baseline algorithms. These results underscore the ability of the algorithm to efficiently and cost-effectively navigate complex obstacle configurations. Beyond its technical merits, this study underscores the broader implications of integrating digital twin technology with path planning algorithms. The digital twin system is integral to the success of the method by enabling seamless data exchange and synchronization, bridging the gap between the physical manipulator and its simulation environment. This ensures that paths generated in the integrated virtual and physical environments are not only theoretically optimal but also practically executable, significantly enhancing the reliability and applicability of the proposed approach in industrial settings. Additionally, the digital twin environment serves as a robust platform for advanced simulations and algorithm refinement prior to real-world implementation. In summary, this paper presents an innovative and robust solution to path planning challenges for manipulators in CNC machining processes. By integrating dynamic elliptical constraints, adaptive step size, and digital twin technology, the proposed method addresses the complexities of irregularly shaped obstacles and dynamic operational environments. The demonstrated reductions in operation time and path cost highlight the method’s potential for widespread application in CNC machining and other industrial automation domains. Future research will focus on extending this approach to multi-manipulator systems, scaling its applicability to larger and more dynamic production environments, and exploring its use in diverse industrial tasks such as assembly, welding, and inspection. This research makes a significant contribution to the development of path planning methodologies within the context of intelligent manufacturing, offering valuable insights for industrial automation solutions.