In this work, we present a methodology for the fast and robust solution of large-scale wave propagation and wave-structure interaction problems. The approach consists in adopting and coupling two different formulations: an Eulerian Shallow Water (SW) approach and a Lagrangian Navier-Stokes formulation. The classical hypotheses of the SW formulation (hydrostatic pressure distribution, negligible vertical velocity compared to the in-plane velocities) allows to solve large three-dimensional problems by practically using a 2D approach, making it particularly useful for simulating wave propagation problems in large domains. When the hypotheses underlying the SW formulation are no longer valid (i.e. in the near-field), the complete Navier-Stokes problem must be solved. Here, for this purpose, we use the Particle Finite Element Method (PFEM), a Lagrangian numerical approach particularly suited for free-surface flows and fluid-structure interaction problems . The two different methods are coupled with a one-way approach. In particular, first, a SW simulation is run, and information in terms of fluid horizontal velocity (averaged on the water depth) and water level is stored; in a second step, the Navier-Stokes problem is solved with the PFEM, using as input the information stored during the SW solution. The proposed algorithm allows a significant reduction in computational costs due to the adoption of the SW formulation in the far-field domain without a loss of accuracy in the near-field zone. The resulting methodology can be employed for different risk assessment analyses, including, for example, the study of coastal water run-up due to sea-storms.