<p>The multi-objective optimal design of rim-driven thrusters (RDT) permanent magnet synchronous motors (PMSM) is a challenging problem due to multiple influencing factors and the low optimization efficiency. This paper proposes a sequential subspace multi-objective optimization strategy based on the intelligent optimization algorithm. The finite element analysis (FEA) and experimental verification of the existing RDT motor are carried out to validate the accuracy of the simulation analysis. According to the Pearson and Spearman’s rank correlation coefficient analysis combined with the design of experiments (DOE) technology, the optimization parameters are divided into two subspaces for optimization. To improve the optimization efficiency, the Kriging approximation model is combined with the non-dominated sorting genetic algorithm (NSGA) in each subspace optimization process to obtain the final optimization model. To validate the accuracy of the approximate model and assess the efficacy of the optimization process, the FEA method was employed to evaluate the optimized motor. A comparative analysis was conducted between the proposed method and the traditional approach, focusing on the obtained Pareto solution sets and computational resource usage. The findings indicate a significant improvement in optimization effectiveness with the proposed method. The optimized RDT motor has a larger average torque and smaller torque ripple and loss compared with the reference motor. The optimization method can effectively reduce the required optimization time and computing resources. This method has a reference value for solving the high-dimensional multi-objective optimization of the RDT motor.</p>