This　paper　presents　a　method　for　optimizing　the　multidisciplinary　shape　design　of　underwater　vehicles　using　a　dynamic　proxy　model.　The　method　employs　a　collaborative　optimization　approach　that　considers　various　disciplines,　including　rapidity,　maneuverability,　energy　consumption,　and　structural　strength　of　the　underwater　vehicle.　The　K　and　T　indices　are　effectively　utilized　to　represent　the　maneuverability　performance　of　underwater　vehicles.　The　hydrodynamics　of　underwater　vehicles　are　analyzed　using　the　Computational　Fluid　Dynamics　(CFD)　numerical　simulation　method.　To　reduce　the　computational　burden　in　the　optimization　loop,　this　paper　proposes　a　dynamic　proxy　model　that　combines　the　trust　region　with　the　adaptive　minimum　confidence　Lowest　Credible　Bound　(LCB)　and　the　Synthetic　Minority　Over-Sampling　Technique　(SMOTE)　algorithm.　Additionally,　an　adaptive　balance　constant　is　introduced　into　the　proxy　model.　The　collaborative　optimization　framework　employs　a　combined　optimization　algorithm　based　on　the　genetic　algorithm　and　Nonlinear　Programming　by　Quadratic　Lagrangian　Programming　(NLPQLP)　algorithm.　The　results　of　applying　this　optimization　strategy　to　the　SUBOFF　model　demonstrate　its　effectiveness　in　optimizing　the　resistance,　mass,　maneuverability,　structural　strength,　and　energy　consumption　of　the　underwater　vehicle.