Objectives This study aims to improve the endurance and performance stability of autonomous underwater vehicles.

Methods Three common rotating bodies were selected as candidates for the AUV’s main body shape. Three models were developed to calculate their volumes and hydrodynamic resistance. The Myring line type was selected as the preferred main body shape of the AUV. The range of parameters for the Myring line equation was then defined. The Latin hypercube experimental design method was used to generate design variable samples and corresponding response values. Based on the sample points, an approximate model was constructed. After the approximate model was established, analyses of fitting accuracy, sensitivity, and response surface were performed. Once the approximate model achieved the required fitting accuracy, a multi-objective optimization model was established for deterministic optimization, and the optimal solution was obtained. Reliability analysis of the optimal solution was then conducted on the optimal solution, followed by robust optimization to enhance its reliability.

Results The results show that after deterministic optimization, the resistance decreased by 5.96% and the volume decreased by 0.78% compared to the initial values. Compared to the deterministic optimization result, robust optimization increased resistance by 0.35% and volume by 0.14%, while achieving a reliability level of 6σ.

Conclusions Both deterministic and robust optimizations reduce the resistance of underwater vehicles, with robust optimization further enhancing the stability of design variables.