<p>The green flexible job shop scheduling problem (GFJSP) has received widespread attention in the context of Industry 5.0. However, the often-overlooked machine deterioration and maintenance during production result in a gap between scheduling plans and their practical applications. This study develops a model for green flexible job shop scheduling that considers machine deterioration and maintenance (GFJSP-DM) and proposes an enhanced two-stage memetic algorithm (ETMA) for its resolution. In the exploration stage, a heuristic hybrid initialization strategy is employed to generate diverse, high-quality individuals; in addition, a potential solution selection strategy aids the model-driven variable neighborhood local search to conduct a more detailed and effective exploration of the objective space. During the optimization stage, the algorithm presents a right-shift energy-saving strategy, designed based on inverse decoding, to evaluate four scenarios of delayed processing, further reducing the total energy consumption of the scheduling plans. Finally, extensive experimental results on test instances demonstrate that ETMA can effectively solve the GFJSP-DM problem.</p>