Hydrodynamic models are important for the design and control of wave energy converters. Traditionally they have been obtained using velocity potential/boundary element type methods, however associated assumptions of in viscid fluid, irrotational flow, small waves and small body motions are a major limitation of this modelling approach, since WECs are designed to operate over wide wave amplitude ranges experiencing large motions, viscous drag and breaking waves. With consideration of the full range of effects, the physics can be described using the Navier-Stokes equations and implemented using computational fluid dynamics (CFD). However, CFD is too computationally expensive for many purposes, for example in initial design optimisation and in the development of control algorithms, where long-time simulations are required. This paper therefore seeks to combine the strengths of both modelling approaches by developing linear parametric models using system identification techniques on CFD generated data. This new approach is demonstrated by identifying the radiation parameters of a linear hydrodynamic model using information from the object’s free decay oscillation simulated in a numerical wave tank (NWT). It is shown to have the ability to produce different representative models for different operational amplitudes; an advantage over the traditional approach which is only representative over small amplitude conditions.