Against the background of sustainable development, distributed shop scheduling systems considering green manufacturing have received increasing attention. This study investigates the energy-efficient integrated distributed hybrid flow shop scheduling problem (EE_IDHFSP) to minimize makespan and total carbon emission (TCE). A Q-learning hyper-heuristic (QLHH) is presented to address EE_IDHFSP. First, a mixed-integer linear programming model for EE_IDHFSP is developed. Second, an improved Q-learning-based high-level strategy is designed to guide six low-level heuristics based on problem features to explore the solution space. Third, an energy-efficient strategy is proposed to reduce the TCE of feasible solutions effectively. To validate the effectiveness of QLHH, extensive experiments and comprehensive comparisons are implemented on 18 test instances. The experimental results show that QLHH outperforms several existing algorithms in solving EE_IDHFSP.