As an important part of the jack-up offshore platform, the spudcan is of the problems of bulky structure and unequal stress distribution. Taken the spudcan of JU2000E jack-up offshore platform as the research object, the finite element model is established by Ansys software. According to the ABS specification, the strength analysis and verification are carried out under the conditions of pre-ballast and storm self-storage. The plate thickness of each structure of the spudcan is used as the optimal design variable, and the sample points are selected by optimal Latin hypercube sampling. Combined with the Radial Basis Function (RBF) neural network algorithm, the surrogate models of the design variables and each response are established, and the multi-objective optimization of the spudcan is carried out based on the second generation non-inferior solution genetic algorithm (NSGA-Ⅱ). The simulation results show that the weight of the optimized spudcan is reduced, the stress is reduced and the distribution is more uniform.