Landslide-generated tsunamis pose substantial risks to human lives, coastal infrastructure, and economic stability, underscoring the urgency of predictive modeling and systematic risk assessment. Leveraging the Smoothed Particle Hydrodynamics (SPH) method, this study establishes a three-dimensional multiphase flow model where the landslide mass is characterized as a Bingham viscoplastic fluid and the ambient water as a Newtonian fluid. Following rigorous validation against laboratory-scale experiments, the calibrated model is applied to simulate the Wangjiashan landslide in China's Baihetan Reservoir region. Simulation results indicate a peak landslide velocity of 12.13 m/s, generating a tsunami wave that propagates to the Xiangbiling resettlement area with a maximum amplitude of 4.94 m—2.5 m above the critical safety threshold. Sensitivity analyses further quantify the interdependencies between initial water level, landslide volume, and resultant wave heights, revealing that wave amplitude increases nonlinearly with decreasing reservoir levels and scales proportionally with landslide volume. Model predictions exhibit strong concordance with empirical formulations and alternative numerical methods, validating its reliability for predicting tsunami hazards and informing evidence-based risk mitigation strategies in reservoir environments.
 

flow-like landslide; landslide-generated wave; wave propagation; multiphase flow model; Smoothed Particle Hydrodynamics