As smart manufacturing advances, the importance of stable equipment operation and cost-effective maintenance has become increasingly evident. Artificial Intelligence (AI) plays a vital role in anomaly detection and predictive maintenance by analyzing real-time sensor data to identify early warning signs of potential equipment failures. This talk explores the application of machine learning, deep learning, and time-series modeling in building intelligent monitoring systems that can detect abnormal behavior and estimate equipment health and remaining useful life. Case Studies from industries such as semiconductor manufacturing and smart factories will be presented, demonstrating how AI can effectively improve equipment availability, reduce maintenance costs, and enhance production resilience.