Time-of-use electricity pricing has become a critical topic in modern energy management, particularly within the manufacturing sector and large-scale industrial production environments. Focusing on distributed flexible job shop scheduling under time-of-use pricing, this study proposes an improved arctic puffin optimization algorithm to minimize total electricity costs and reduce the production costs of enterprises. First, the latin hypercube sampling is employed to gain a more diverse and uniformly distributed initial population. Second, a dynamic parameter adjustment strategy based on Lévy flight is employed to achieve a balance between the global exploration and local exploitation phases of the algorithm. Next, the ranked order value decoding method is employed to choose low-power processable machines. Finally, a delay processing strategy for the job is proposed to address the scheduling under time-of-use electricity pricing constraints. The results of comparison experiments and industrial case studies demonstrate that this method exhibits excellent performance and stability in solving the distributed workshop scheduling problem under time-of-use electricity pricing constraints.