Abstract: In minimally invasive surgery, accurate position information of the instrument end is indispensable to provide a high control accuracy, and a satisfactory consistency of hand-eye coordination control for surgeons, even if sensors are not installed at the end of surgical instrument. To address this issue, a transmission model of wire rope driven surgical instrument is established based on the principle of the improved Preisach model. The model parameters are fitted using NSGA-Ⅱ (nondominated sorting genetic algorithm Ⅱ), a multi-objective genetic algorithm. Thus, motion compensation for surgical instruments based on the proposed model is achieved. Firstly, an experiment platform of the wire rope drive for minimally invasive surgical instrument is built. Subsequently, motor drive information and end-effector movement information are collected. Finally, model identification and verification are carried out, and the resulting maximum error of the model is less than 1°. To evaluate the algorithm effectiveness, it is compared with the methods of linear regression, random forest, and neural networks. The proposed algorithm demonstrates the best comprehensive index, exhibiting high precision, stable results, and low overfitting.