With the development of virtual reality technology, medical simulation training systems with visual-haptic feedback have gained widespread attention in minimally invasive surgical education and skill training. In laparoscopic procedures, which require precise manipulation of soft tissues, the realistic simulation of interaction process with tissue objects is essential for improving trainees’ surgical skills.  However, the complexity of the nonlinear mechanical behavior of soft tissues makes it difficult to accurately replicate tissue-instrument interactions via both visual and haptic channels. This remains a technical challenge in virtual surgical simulation. To reduce the perceptual discrepancy between virtual training and real environments, this work proposes a psychophysically optimized haptic rendering approach for high-fidelity soft-tissue interaction. In this approach, the real-time deformation of soft tissues is achieved using the Extended Position-Based Dynamics (XPBD) algorithm, while the Kelvin Boltzmann (KB) model is employed to accurately describe viscoelastic mechanical responses between surgical instruments and soft tissues. Additionally, a psychophysical approach is employed to map the relationship between perceived haptic forces in virtual and physical environments. This mapping serves to optimize haptic output, thereby enhancing realism during soft-tissue interactions. Perceptual accuracy test and user experience evaluations were conducted, and the results demonstrated a significant reduction in perceptual inconsistency between virtual and physical interactions, thereby enhancing haptic fidelity in the interaction.