<p>Effective production logistics (PL) delivery is vital for enhancing productivity and reducing costs in discrete manufacturing workshops. However, the dynamic and uncertain nature of the delivery task (DT) poses challenges to the allocation of the DTs. To effectively utilize the real-time PL data and achieve efficient PL delivery for dynamic DTs in discrete manufacturing workshops using industrial internet of things (IIoT) technology, this study proposes a self-organized response method based on non-cooperative game theory for PL delivery in workshops. First, a data paradigm is defined that can reflect the real-time status information of DTs and delivery execution units (AGVs) in the IIoT environment of the discrete manufacturing workshops. Driven by the real-time data of DTs and AGVs, a dual-task sequence rolling self-organized response strategy is proposed, incorporating a partial task delay response rule to mitigate the uncertainty inherent in the DTs. Then, a non-cooperative game-theory-based multi-agent PL delivery self-organized response model is established by integrating time and distance as optimization objectives. An algorithm is designed to obtain the Nash equilibrium solution of the response model as the fully reactive response solution. Finally, the discrete manufacturing workshop of a printing and packaging enterprise was selected as the case study scenario to verify the proposed method. The results show that the method can not only effectively reduce the ineffective transportation distance of AGVs and improve response accuracy, but also further balance AGV utilization.</p>