The escalating demand in global energy requirements and the soaring price of traditional fossil fuels, in combination with global awareness to follow a pathway toward decarbonisation, are fuelling research and development into novel technologies to harness renewable energy sources. Ocean wave energy, a significant and untapped source of renewable energy, if economically viable, can make a promising avenue for sustainable energy generation. In the drive for the development and more efficient operation of wave energy converters (WECs), effective control systems, which can maximise converted energy for a given capital cost, are crucial. Model-based control systems contribute to the majority of energy-maximising control systems of WECs, with a need for a suitable mathematical model. Physics-based models, numerical simulations, full-scale tests, or laboratory-scale tests can be used to develop WEC models, each presenting distinct methodologies and challenges, yielding models with a diverse range of accuracy. The effectiveness of model-based control relies on the precision of the WEC model upon which the controller is based, given that WEC controllers have shown considerable sensitivity to inaccuracies in their underlying models. Among the distinct WEC modelling techniques, the determination of models from data (using either physical or numerical experiments) is an effective route to derive representative (linear and nonlinear) WEC models. In this thesis, both numerical wave tank (NWT) and physical wave tank (PWT) experiments are considered to estimate a range of adequate linear WEC models capable of meeting the requirements of a control system by the employment of the proper test signals (specific to each test setup) to cover the full operational space of the system. Many uncertainty parameters in the data-driven models with sources differing significantly in NWT and PWT experiments, may hinder accurate WEC model determination for efficient (model-based) control strategies. Within this scope, the current study considers the range of tests that can be performed, the uncertainty sources, and the range of post-processing techniques that can be applied in NWT and PWT tests for a point-absorber type system, with an aim to get the most realistic hydrodynamic WEC models. Moreover, in NWT testing, a comprehensive representation of input signal synthesis and characterisation is carried out, to provide sufficient fidelity in Computational Fluid Dynamics (CFD)- based NWT tests and, in PWT testing, specific focus is directed towards the quantification of uncertainty and external disturbance, specifically tailored to the PWT under study, with the corresponding effects are mitigated by applying effective data-processing steps. Multi-linear hydrodynamic WEC models (obtained either from NWT or PWT testing), serve as a starting point for model-based linear WEC controller synthesis. Finally, with a view to robust WEC control, a (single) nominal model and uncertainty bound are quantified from multi-linear models obtained from NWT tests, and robust control results are provided to demonstrate the efficacy of the modelling and control philosophy.