The vertical revetment is a commonly used structural form in coastal engineering.There are numerous factors affecting wave overtopping under wave action，especially considering the actual multi-directional wave action，and currently，there is no unified calculation method.This paper utilizes the advantages of neural network methods in regression problems and constructs a BP neural network model.By using physical experimental data to build the training set，a predictive model for wave overtopping of sloping revetments under multidirectional wave action is established.Through comparisons with the test set，empirical formulas，and other experimental results，the model demonstrates excellent prediction accuracy and can satisfy engineering design needs.Furthermore，the model is utilized to analyze the influence of wave incident angle and directional spreading on wave overtopping，addressing the issue of indistinct variations caused by incomplete experimental data