Through physical modeling experiments，we study the effects of the mean breaking wave parameter，the relative width of the shoulder platform，the relative height of the wave wall and the relative height of the dike on wave overtopping of the seawall under the action of irregular waves，apply the artificial intelligence algorithm to establish the predictive model for wave overtopping，and verify the model combined with the actual project. The results show that wave overtopping increases with the increase of the mean breaking wave parameter and decreases with the increase of the other three influencing parameters. The wave overtopping has a complex nonlinear function relationship with these four factors. The artificial intelligence algorithm model has good predictive accuracy and generalized performance. The prediction errors of the over 90% data in the test group are within ±10%.