Detecting and identifying anomalies that cause significant deviations from expected supply chain operations is important for maintaining efficiency and preventing major disruptions in the flow of materials and products. The identification and addressing of these anomalies allows companies to ensure smooth operations, minimize delays, reduce costs, and improve supply chain performance. Thus, effective anomaly detection helps to mitigate risks and maintain the reliability of supply chain processes. This chapter focuses on anomaly detection in the context of smart manufacturing through the use of an Intelligent Decision Support System. The proposed system consists of a prediction subsystem that, in turn, feeds a system dynamics model that simulates the states of the industrial plant. Additionally, it uses a deep learning-based network with an attention mechanism to detect disruptions. In simulation studies, the network is compared against other methods, demonstrating superior performance in the task of disruption detection. Thus our approach not only enhances the accuracy of anomaly detection but also could improve the efficiency and resilience of the supply chain management process.