针对弱支撑环境下近海渔场微网群光伏出力与负荷预测难以获得可靠概率分布，导致互联微网群能量调度中的经济性问题，开展计及源荷不确定性的优化运行研究。从渔业生产过程出发，建立单渔场微网结构与关键设备模型，并通过联络线构建互联微网群，建立考虑联络线容量与传输效率的确定性调度模型；在此基础上引入信息间隙决策理论（Information-Gap Decision Theory，IGDT），分别构建鲁棒型模型以最大化可承受不确定性半径刻画韧性边界，构建机会型模型以最小化达成收益目标所需的不确定性半径挖掘潜在收益空间，从而形成面向不同风险偏好的统一决策框架。算例表明，能量互济协同可降低渔场微网群综合运行成本：离网能量互济相对孤岛运行成本下降约 3.9%，混合并网模式下降约 54.8%；同时，鲁棒型策略在一定经济性让渡下显著提升抗扰能力，并抑制柴油机频繁启停，机会型策略则在可控风险下获得更优经济性表现。本文方法能够在不依赖精确概率分布的条件下实现互联微网群运行经济性的定量权衡，为弱支撑场景的优化运行提供可解释的决策方法。

In offshore fishery microgrid clusters operating under weak support, reliable probability distributions are often unavailable for PV generation and load forecasts, which challenges economic energy scheduling. This paper studies optimal operation considering source–load uncertainty. Starting from the fishery production process, a single-farm microgrid structure and key component models are established, and an interconnected microgrid cluster is formed via tie-lines. A deterministic dispatch model is developed considering tie-line capacity and transmission efficiency. Information-Gap Decision Theory (IGDT) is then introduced to build a robustness-oriented model that maximizes the admissible uncertainty radius to characterize the resilience boundary, and an opportunityoriented model that minimizes the uncertainty radius required to meet a target benefit to explore potential gains, providing a unified decision-making framework for different risk preferences. Case studies show that energy mutual aid reduces the total operating cost: compared with isolated operation, the cost decreases by about 3.9% under islanded mutual aid and by about 54.8% under a hybrid grid-connected mode. Moreover, the robustness-oriented strategy significantly enhances disturbance tolerance at the expense of limited economic performance and suppresses frequent diesel generator start–stop actions, while the opportunity-oriented strategy achieves better economic performance under controllable risk. The proposed method enables quantitative economic trade-offs for interconnected microgrid clusters without relying on precise probability distributions, and provides an interpretable decision-making approach for offshore fishery microgrid clusters and other weak-support scenarios.