Prediction of wave overtopping rate is an important step in the functional design of vertical coastal structures. In this study, scaling arguments and data mining approaches were used to derive formulae for the prediction of wave overtopping rate at fully vertical and nearly vertical impermeable solid smooth structures. An extensive database mainly selected from the CLASH including impulsive and non-impulsive waves, low and high freeboard and composite structures were used for formulae development. The obtained dimensionless overtopping rates were compared with those of existing formulae. The performances of formulae were first assessed quantitatively using laboratory tests mainly within the CLASH database; and it was shown that the developed formulae are more accurate than the previous ones. In addition, the developed formulae were applied to the field measurements. Results indicated that the developed formulae which consider the effects of all governing parameters, perform better than the previous empirical formulae with at least 18% and 31% RMSE improvement in predicting the overtopping rate in laboratory and field experiments, respectively.