Reliable prediction of wave overtopping rate is a key task in the design and safety assessment of coastal structures. The overtopping rates must be lower than the allowable rate both in normal operating and extreme conditions to guarantee the safety of both people and assets on and behind the breakwaters. The sea level rise caused by climate change and its effects on wave climate enhance the overtopping rate and make the existing coastal structures more vulnerable to the overtopping. In the past decades, several approaches have been developed for prediction of overtopping rate in different structures, including empirical formulae, numerical models, and soft-computing methods. The most commonly used models are empirical ones. Typically, these models are developed by fitting dimensionless parameters to experimental data obtained from physical model tests. The most popular empirical formulae proposed for berm breakwaters are the formulae of Van der Meer and Janssen (1995) and EurOtop (Pullen et al. 2007). However, the applicability of these formulae is limited to specific structures and wave conditions for which the models were developed. Hence, a more comprehensive dataset which includes multiple tests with different structural and hydraulic conditions would enhance the robustness and reliability of the developed models. As a part of the European CLASH project framework, two different ANN models for various types of coastal structures are developed. These models can also be used for the prediction of wave overtopping rate at rubble mound breakwaters with berms. However, using ANN models may not always be desirable, because, they have complicated structures and are not as transparent and understandable as empirical models. On the other hand, practical applications of these models need engineers to have some data mining knowledge which is not always the case. In this study, a new formula for the prediction of wave overtopping rate at non-reshaping berm breakwaters is presented. In order to derive the new formula, extracted experimental data from the CLASH database is implemented and the effects of the most important governing parameters are considered. Several models were developed using various dimensionless parameters that seemed to be influential and the model which outperformed the other ones was selected. The final model considers the effects of dimensionless crest freeboard, dimensionless crest width and average angle of the structure slopes. The new formula’s performance is then compared with those of previous empirical ones for prediction of overtopping rate. Statistical indicators such as Root Mean Square Error, RMSE show that the new formula outperforms the previous empirical ones.