ObjectiveSAiming at the problems that traditional surgical training relies on in vitro models and the contradiction between simulation authenticity and computational real-time performance of virtual surgery systems (VSS), this study focuses on the three core technologies of geometric modeling, deformation simulation, and haptic force feedback for virtual surgery on human liver soft tissue, so as to provide technical support for the clinical transformation of VSS.SMethodsSThe MA-UNet segmentation network integrating multi-scale spatial attention mechanism and dynamic dilated convolution was proposed, and the high-precision tetrahedral mesh model of the liver was constructed combined with Delaunay triangulation. Extended position-based dynamics (XPBD) was adopted and a small-step iteration strategy was introduced to balance the authenticity and real-time performance of deformation. A hybrid spring model coupled with finite element (FE) parameters was proposed to optimize the accuracy of instrument-tissue interaction force feedback. Based on the Decathlon liver dataset (131 cases of CT images), the technical performance was verified by indicators such as Dice coefficient, volume error, and mechanical error.SResultsSIn the liver segmentation task, the MA-UNet achieved a Dice coefficient of 94% and an intersection over union (IoU) of 86%. The small-step XPBD realized real-time deformation with a Dice coefficient of 97%, a time consumption of 21.3 milliseconds per frame, and a volume error of only 2.2%. The hybrid spring model had a mechanical error of 0.21 N and a calculation time of 1.7 ms/frame, and the coincidence degree between the stress-strain curve and the in vivo measured data was ≥95%.SConclusionSThe proposed technical system can effectively solve the problems of inaccurate segmentation, distorted deformation, and untrue feedback in virtual surgery of liver soft tissue. It provides a high-fidelity simulation platform for surgical training and preoperative planning, and is of great significance for promoting the clinical application of VSS and improving the safety of surgical operations.