An improved Analytic Hierarchy Process-Criteria Importance Though Intercriteria Correlation (AHP-CRITIC) subjective and objective weighting method is proposed for the optimization of the main dimensions of semi-submersible platform. Based on the response surface optimization method, the functional relationship between design variables and objective function is obtained. The Multi-Objective Genetic Algorithm (MOGA) is used to optimize the dimensions of the main components of the 7th generation semi-submersible platform. The Pareto front and three sets of optimal solutions are obtained. Compared with the original scheme, the hydrodynamic performance is significantly improved and the overall steel weight is significantly reduced. The original single subjective and single objective decision-making methods are combined and improved by AHP-CRITIC combined method to evaluate three different sets of the optimal solutions, which can satisfy both the subjective expectation of decision maker and the laws of correlation among the indexes. Compared with the single decision-making method, the combined method is more comprehensive and reasonable, and can be applied to determine the main dimensions of the preliminary design of practical projects.