Analysis of thermodynamic parameters’ effects on the S-CO2 recompression Brayton cycle for recovering the main engine exhaust gas waste heat of an ocean-going 9000 TEU container ship are carried out first in this paper. NSGA-II algorithm-based multi-objective optimizations are conducted to maximise net power output and exergetic efficiency as well as to minimise the levelized cost of energy (LCOE). The final optimal result from Pareto front solutions is decided by decision-making processes. The results show that the LCOE increases along with the enhancement of net power output and exergetic efficiency, and the maximal objectives always appear near the upper boundary of exhaust gas temperature. The final optimal result decided by TOPSIS decision-making has higher rationality and accuracy than that obtained by the LINMAP method. Regarding the final optimal results of triple-objective and dual-objective optimizations obtained by the TOPSIS method, the former achieved more reasonable results, since its optimal net power output and exergetic efficiency are 1.29% and 5.11% higher than the latter, while its LCOE increased by 2. 28% as a result of the increase of the net power output. The recompression Brayton cycle is better for recovering the exhaust gas waste heat in ships compared with the simple recuperative cycle without recompression.