The development of effective energy-maximising control strategies has a crucial role in the empowerment of wave energy technology, and in its improvement towards economic viability. Within the state-of-the-art, most of the strategies adopted to maximise the absorbed energy exploit a model of the wave energy converter (WEC) to be controlled, i.e. they are model-based. These models attempt to replicate the WEC dynamics with a sufficient degree of fidelity, trying, at the same time, to minimise their associated computational burden. However, due to the presence of the hydrodynamic effects , which inherently characterise wave energy systems, simultaneously achieving high-fidelity and computational efficiency is not trivial. Oversimplification of the problem through, for example, linearity assumptions, could lead to non-representative models and/or large uncertainty levels. To overcome these issues, in the last decade, several approaches based on data have been proposed in the wave energy field. These approaches, falling under the umbrella of system identification techniques, exploit data coming from experimental tests or high fidelity simulations, and build control -oriented models with a pre-defined level of complexity. In this paper, we analyse the different strategies that have been adopted in the literature to build data-based control-oriented models for WECs, highlighting the characteristics of each approach, together with their opportunities and inherent drawbacks. An analysis of eventual “partial” data-based modelling of WEC subsystems (e.g. moorings, PTO, or hydrodynamics only) is also reported. Moreover, considerations on the choice of inputs and outputs depending on the WEC type are reported, in an attempt to highlight the different issues that characterise the system identification problem depending on the WEC technology. Finally, conclusions are drawn regarding the capabilities that this type of approach has in (at least partially) solving the modelling issues that affect WEC control system design, and the pitfalls that pure adoption of these strategies has when applied on larger scales, or in the operational stage.