Abstract <p>This review depicts the most recent developments in sensor-based machine learning approaches for the identification and prediction of surface defects in machining operations, which are essential for ensuring accuracy, quality, and manufacturing productivity. The review spans the period from 2018 to 2024, and the major emphasis is on the employment of a variety of sensors, such as vibration, cutting force, optical, laser triangulation, acoustic emission, and capacitance, along with machine learning (ML) and deep learning models for the accurate prediction of surface roughness. Machine learning and deep learning algorithms, including convolutional neural network (CNN), support vector machine (SVM), long short-term memory (LSTM), gated recurrent unit (GRU), and combinations of these algorithms, like CNN-SVM and CNN-GRU, that greatly improve prediction accuracy, as demonstrated through a systematic review of 113 works of research. The integration of these models with live or almost-live data collection systems renders them very flexible and strong against changes in manufacturing conditions. The review also discusses signal preprocessing and sensor data fusion used to enhance prediction performance and provide adaptive feedback control to optimize processes. Nevertheless, these models are still effective, but they have certain problems that restrict their use in more complex industrial settings that are noisy, variable, and lack real-time data to generalize the models. Machine learning techniques enabled by sensors, especially hybrid frameworks, are key to the performance optimization of milling, the quality improvement of manufacturing, and the evolution of intelligent, green production systems.</p> Graphical Abstract <p></p>