Abstract: In view of the high potential risk, limited vision and poor consistency of the traditional surgical approaches to cochlear implantation, a multi-constraint and multi-objective preoperative planning optimization algorithm is proposed, which aims at the shortest DCA (direct cochlear access) from incision to round window and the least damage during implantation. Firstly, the operation and image integrated surgery path planning for robotic cochlear precise implantation is studied. The reconstructed 3D surgical view with multi-modal information fusion is taken as the working space. The identification of cochlear and facial nerve is carried out based on U-Net segmentation and recognition algorithm, which can provide vectorized functional partition and boundary information. The multi-objective and multi-constraint planning algorithm based on 3D reconstruction of CT images in Cartesian coordinate system is adopted. According to the design rules of robotic cochlear implantation surgery approach, the formula of path planning algorithm is deduced, and the corresponding DCA path is planned for the manipulator. Through the simulation experiment on the virtual simulation platform, the effectiveness and accuracy of the proposed method are proved.