Abstract: Renewable energy has been one of the important solutions to the current energy crisis and environmental pollution. The low-grade-energy driving technology has also been widely used to improve energy efficiency in the fresh water production, with the increasing water demand in recent years. Among them, the distributed cogeneration system using Proton Exchange Membrane Fuel Cell (PEMFC) can be expected to efficiently provide the multiple energy demands and freshwater. The waste heat of PEMFC can then be used to directly drive the membrane distillation system, since the operating temperature of Direct Contact Membrane Distillation (DCMD) is around 40-70℃, which is matching with that of PEMFC around 60-80℃. In this study, a novel distributed cogeneration system was proposed to recover the PEMFC waste heat using DCMD for the drinking water and high energy efficiency. The mass and heat transfer models of DCMD, and the PEMFC electrochemical-thermodynamic model were established to verify the experimental data. The results demonstrated that the model prediction was in better agreement with the experimental data, indicating the model suitable for further analysis. A mathematical model was established to demonstrate the cogeneration system performance under design and off-design conditions. It was found that the power generation efficiency, transmembrane water flux, the overall energy efficiency, and exergy efficiency were 19.26 kW, 15.340 5 kg/(m2•h), 57.77%, and 52.68%, respectively, under the design condition. The overall energy efficiency of the cogeneration system increased by 12.91% in the PEMFC and DCMD system, compared with the simple PEMFC system. In the case of the off-design condition, the hydrogen stoichiometry and the mass flow rate of permeate water can greatly contribute to consume consuming more extra electric energy for a higher DCMD membrane flux. The oxygen stoichiometry was contributed to improve improving the overall exergy efficiency in the cogeneration system via reducing exergy loss of PEMFC. However, the temperature at the feed inlet posed a significant effect on the DCMD performance, and the heat exchanger 3 and DCMD presented a small proportion of exergy loss coefficient. A genetic Non-dominated Sorting Genetic Algorithm II (NSGA-Ⅱ) was carried out to obtain the optimal operating conditions of the PEMFC/DCMD cogeneration system, due to the low computational complexity, high running speed, and excellent robustness. The decision variables were set as the hydrogen and oxygen stoichiometry, permeate water mass flow rate, and pinch point temperature in the heat exchanger 3. The optimization objective parameters were taken as the power generation efficiency, transmembrane water flux, the overall energy efficiency, and exergy efficiency. As such, two schemes were finally divided to optimize. In addition, the Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) was used to determine the optimal design conditions of the system under different weights for each objective. The 3D Pareto frontier was selected in the multi-objective optimization of the cogeneration system. It was found that the maximum of DCMD membrane flux, the overall energy efficiency, and energy efficiency increased by 9.19%, 2.48%, and 4.78%, respectively, compared with the system performance under the design designed condition conditions. The excellent performance of the PEMFC/DCMD cogeneration system can be widely expected to apply for to the PEMFC/DCMD based electricity and fresh water cogeneration system.