Abstract: For the multi-factor flexible job shop green scheduling problem that integrates machine variable-speed and worker assignment, we establish a mixed-integer programming model with objectives to minimize makespan, total machine load, and electricity cost, and propose an improved decomposition-based multi-objective memetic algorithm. Firstly, we design a multi-layer chromosome encoding mechanism to represent four sub-problems: Operation sequencing, machine selection, speed level, and worker assignment, and use a hybrid strategy to generate a high-quality initial population. Secondly, we develop an adaptive local search strategy based on problem characteristics, enabling the adaptive selection of local intensification operators and dynamic allocation of computational resources. By combining cooperative evolution with adaptive local search, the algorithm balances global exploitation and local exploration capabilities. Finally, simulation experiments based on extended benchmark instances and a real-world case demonstrate that the proposed algorithm significantly outperforms four other state-of-the-art algorithms, validating its effectiveness and superiority in solving multi-factor green scheduling problems.