BACKGROUND: In an unstructured environment where real-time human decision is essential, shared control allows collaboration between humans and robotic systems, combining advantages of both. However, existing control methods are challenged with precision loss, inconsistency and interference from unconscious human inputs. METHODS: An adaptive anisotropic control frame is presented, enabling interaction both operational and tactical levels. Using predefined trajectory, a dynamic weight function is proposed to allow the human operator to override. Movement along preferred direction is encouraged and compensated, providing accurate real-time tracking performance. Haptic feedback during shared control is evaluated and optimised. RESULTS: Experiments validate that the raised method can achieve a tracking precision of CONCLUSION: The proposed approach provides both stability and flexibility in interactive surgical manipulations, maintaining similar precision with autonomous execution.