Abstract: To address the non-uniform airflow distribution caused by asymmetric air duct layouts in containerized energy storage systems, a staged air duct structure optimization strategy based on response surface methodology (RSM) is proposed in this paper. Firstly, the height and width of air duct openings are set as design variables. Combined with simulations, response surface models for airflow non-uniformity and flow resistance are established, and an initial optimal configuration is obtained via multi-objective optimization. Subsequently, flow-guiding vanes are mounted at four air outlets, and the length and installation angle of the vanes are optimized through a second-stage RSM iteration. Finally, a comprehensive integrated duct performance index (IDPI) is introduced to quantitatively evaluate the design alternatives. The results demonstrate that the optimized airflow non-uniformity is reduced to 0.06, the IDPI is significantly improved, and the temperature uniformity among battery clusters is obviously enhanced. Simulation results indicate that the maximum temperature difference of the dual-cluster model is limited to 4.5 °C and the maximum temperature rise is controlled below 8.5 °C. Test data under dynamic variable operating conditions show that the maximum temperature difference is approximately 5.0 °C and the maximum temperature rise is 8.0 °C. The effectiveness and engineering feasibility of the proposed asymmetric duct structure with the “top-side air supply and front-side return” mode and the terminal regulation strategy are verified under complex operating conditions.