The importance of properly modelling the effects of air compressibility in the selection of the optimal design parameters for an Oscillating Water Column wave energy converter is investigated. For this purpose, a wide dataset of capture width ratios, obtained from both experimental tests and Computational Fluid Dynamic simulations, is used to formulate an empirical model able to predict the performance of the device as a function of its basic design parameters (chamber width and draught, turbine damping) and of the wave conditions (wave period, wave height). A multiple non-linear regression approach is used to determine the model numerical coefficients. The data used to formulate the model include the effects of air compressibility. The impact of considering such effects on the selection of the optimal geometry of the device is evaluated and discussed by means of the model application for the optimization of a device to be installed in a site located in the Mediterranean Sea (in front of the coast of Tuscany, Italy).