The interaction between ice, water, and structures in northern rivers and seas plays a critical role in planetary processes and human activities. These systems involve highly complex multiphysics phenomena, including large deformations, fracture, multi-body interactions, and strong solid–fluid coupling, which make their numerical modeling and prediction particularly challenging. Particle-based methods offer a powerful and flexible framework for addressing these challenges. In this work, we present particle-based approaches for simulating ice–water dynamics using both discrete and continuum descriptions of ice mechanics. In the discrete-based approach, ice is modeled as an assembly of interacting floes representing a brittle or quasi-brittle solid. For this framework, we develop an SPH-based fluid–structure interaction (FSI) model capable of explicitly resolving ice deformation, fracture, floe formation, and post-failure multi-body contact. In contrast, the continuum-based approach treats ice as a continuous material field governed by constitutive laws. We present a multiphase SPH/MPS formulation with viscoplastic ice rheology, for both fully three-dimensional and two-dimensional depth-averaged simulations of ice dynamics. The proposed models are validated against a range of benchmark problems, including buoyant granular collapse, wave–ice interaction, and river ice jam formation, with comparisons to available experimental measurements. The results demonstrate that both particle-based approaches accurately capture key features of ice–water interaction while offering complementary strengths in terms of simulation scale, numerical robustness, and computational efficiency.