Mobile boats hoists (MBH) are huge mobile cranes that are used to carry floating from water to land for maintenance. Floating's weights are so high and so buckling in both global and local modes is a big challenge in design of the MBH structure. According to the crane design standards, the stability factor of safety of the crane's structures must be at least 1.34.The main goal of this dissertation is to optimize the design of the beams and columns of an 800 tons capacity MBH to minimize the weight of its structure relying on preventing buckling in the beams and columns. Solution used to prevent buckling, especially local buckling, is to improvise intermediate stiffeners with U cross sections in the beams and columns. The length, width and thickness of the intermediate stiffeners improvised in the beams and columns are considered as the design parameters. The weight of the structure is assumed as the objective function of optimization process. The optimization constraint is to avoid buckling according to the standards. To achieve the optimization goal, finite element model of the structure with shell elements is produced in ABAQUS. The produced model is parametric according to the length, width and the thickness of the intermediate stiffeners. In the following, Taguchi method is used for design of experiments and requirement data are extracted. In the Taguchi method according to conditions of issue, the L' 36 orthogonal array is selected. After design of experiment using Taguchi method, the extracted data are converted to mathematical formulation using neural network. To do this, a two-layer back propagation neural network with hyperbolic tangent sigmoid transfer function for first layer and linear transfer function for second layer is used. Finally using genetic algorithms the optimum weight and value of design parameters in optimum point are gained. The genetic algorithm process is don using MATLAB toolbox. This optimization process shows that in the bottom longitudinal beams and in the front and rear columns of the structure, no local buckling happens and then no intermediate stiffeners is needed for this beams and columns. At the end, using values gained from genetic algorithm and using parametric model, the weight and the stability factor of safety of structure are calculated and the results show that constraint of the problem is satisfied with the stability safety factor of 1.35. Keywords: Optimization, Buckling, Stiffener, Finite Element Model, Neural Network, Genetic Algorithm