Plug-in photovoltaics (PV)/battery power systems attract increasing interest due to their merits in zero fuel consumption and zero local emission. However, current optimization for the plug-in PV/battery power system only considers a single objective of minimizing lifecycle cost, which may result in increased lifecycle greenhouse gas (GHG) emission due to electricity generation from coal in some areas. The paper therefore proposes a bi-objective optimization methodology to find out an optimal trade-off between lifecycle cost and GHG emission. Non-dominated sorting genetic algorithm II is developed to explore the Pareto solution sets. An unmanned patrol boat is considered as a study case. Simulation results show that the optimal design from the bi-objective optimization gains a 12.6% reduction in lifecycle cost and a 53.8% reduction in GHG emission when compared with the conventional power system consisting of diesel engines and generators. Moreover, the optimal design achieves a 46.3% reduction in GHG emission compared with the single-objective one aimed at minimum lifecycle cost, which at the same time increases the lifecycle cost by just 0.6%. Besides, two variables (the number of PV array and the number of battery modules) are found to be the most sensitive to the contradiction between lifecycle cost and GHG emission.