Abstract: To ensure the normal operation of radial gates under sediment deposition conditions, accurate prediction on their maximum opening force is essential.First, using the meshless particle method in Particleworks to simulate sediment deposition environments, we constructed a multi-domain coupling simulation model integrating mechanical, hydraulic, and Bingham fluid dynamics.This model generated extensive simulation data.Next, a BP neural network surrogate model, combined with the whale optimization algorithm (WOA) and kernel density estimation (KDE), was employed to train the simulation data, enabling rapid point prediction and interval prediction of the opening force.The results demonstrated that the multi-domain coupling simulation model predicted the maximum opening force with an error margin within 4% compared to physical experimental data.The WOA-BP-KDE neural network model achieved a prediction error less than 3% for simulation samples, with satisfactory interval prediction performance.These findings provide valuable insights for simulating the influence mechanism of sediment deposition on gate opening force and improving prediction accuracy.