Abstract: To address the optimization problem of energy-efficient automated assembly line configuration, an optimization model is established with the objectives of minimizing total production cost, equilibrium index and production energy consumption, while comprehensively considering process sequences, workstation parallelism, and equipment power. An improved multi-objective harmony search algorithm is proposed to solve the problem. A random key encoding/decoding method based on the process priority relationship is designed. Considering the reliance of the harmony search algorithm on the initial harmony memory library, an elite library generated by fast non-dominated sorting is used to initialize the harmony memory library to improve the quality of solutions in the harmony library. Additionally, a variable neighborhood search strategy is embedded to improve the local search capability of the algorithm. By comparing with three algorithms, it is verified that the proposed method has better solutions with more evenly distributed Pareto fronts. The proposed method is applied to a case study of a fully automated assembly line for smart meters, resulting in nine configuration schemes, which meet the different needs of enterprises and verify the effectiveness of the proposed method.