<p>In the field of mechanical processing, model prediction and manufacturing process optimization are shifting from experience-driven and trial-and-error-based approaches to data-driven and artificial intelligence (AI)-enabled methodologies. In recent years, AI technologies have achieved promising advancements in parameter prediction, condition monitoring, and multi-objective optimization, emerging as a pivotal driver in the evolution of manufacturing practices. AI advancements and practical applications over the past decade in typical machining scenarios (cutting, grinding, lapping/polishing), and non-traditional machining were reviewed, categorized into three core tasks of prediction, monitoring, and optimization. This review summarizes typical integration pathways for various AI methods in multi-source data fusion, physical constraint modeling, and real-time control, analyzing strengths, weaknesses, and applicability under complex operating conditions. Despite substantial progress, several challenges persist, including difficulties in integrating data with physical mechanisms, insufficient control and decision-making capabilities under dynamic conditions, and limited interpretability of deep learning models. Emerging trends in integrating physical mechanisms, transfer learning, and explainable AI are further explored, aiming to provide theoretical foundations and methodological references for intelligent mechanical processing.</p>