In the modern manufacturing environment, the production systems develop towards a flexible direction due to the increasing of the market demands of multiple varieties and small batch-based customized products. In order to make better use of existing resources and improve production efficiency, real-time performance evaluation and prediction of real-time scheduling and optimization improvement based on small batch production have important research significance in distributed flexible production systems. This paper mainly studies the performance analysis of multi-batch serial production lines based on degenerate machines, and carries out task scheduling and predictive maintenance for the distributed production system. Specifically, for the production system with the degenerate machine model and the finite capacity buffer, the mathematical model is established using the Markov method firstly. Then, an accurate method is proposed to calculate the real-time performance indexes of the system based on the model, and an algorithm is designed to optimize the completion time index for the scheduling problem under the model. In addition, a predictive maintenance strategy is proposed to reduce the completion time for machines with degradation models. Finally, the feasibility and effectiveness of the mathematical model and algorithm are illustrated by numerical experiments.