Abstract: Highly viscous non-Newtonian free-surface flows are widely encountered in industrial applications, where the complex rheological properties and evolving interfaces present significant challenges for numerical simulation. The Smoothed Particle Hydrodynamics (SPH) method, as a representative mesh-free computational approach, offers unique advantages in handling free-surface flows. However, the numerical accuracy of the physical viscosity term in SPH is highly sensitive to the discretization scheme and particle distribution, with errors becoming particularly pronounced in high-viscosity flow scenarios. Meanwhile, the stringent timestep constraints imposed by high viscosity severely reduce computational efficiency. To address these issues, this study employs the non-Newtonian Cross model to characterize the fluid viscosity behavior, and integrates a robust Particle Shifting Technique (PST) together to improve particle distribution uniformity. Based on two typical viscous formulations, the MGF and MEA forms, both explicit and implicit viscosity SPH numerical models are developed. The explicit viscosity SPH model is suited for simulating free-surface flows of low- to medium-viscosity non-Newtonian fluids, whereas the implicit viscosity SPH model is designed for high-viscosity flow scenarios, enabling significant improvements in computational efficiency and stability under stringent timestep constraints. Systematic validation is performed using canonical high-shear flow cases, including droplet impact, jet impingement, and jet buckling. The results demonstrate that the MGF viscosity formulation consistently maintains good stability and physical consistency in both explicit and implicit viscosity SPH models, supporting high-fidelity simulation of high-viscosity, high-shear flows. The implicit solution strategy substantially enhances computational efficiency under high-viscosity conditions, providing an effective numerical approach for efficient, stable, and physically faithful simulation of non-Newtonian flows in complex scenarios.