<p>The combination of cutting-edge sensor technologies and AI-based automation has led to notable advancements in intelligent welding systems in recent years. With an emphasis on thermal imaging for real-time weld quality assessment, optical tracking systems, and sound-based monitoring, this thorough review examines the state-of-the-art in intelligent welding. Accurate seam tracking, joint geometry detection, and adaptive control are made possible by optical systems that make use of cameras, laser triangulation, and structured light. In order to detect flaws such as porosity, spatter, and incomplete fusion, sound-based monitoring records acoustic emissions. Understanding temperature dynamics, melt pool behavior, and penetration depth is possible through thermal imaging and heat flow mapping. To improve defect detection accuracy, process interpretability, and adaptability to changing welding conditions, the review emphasizes the integration of these sensing modalities with AI techniques, such as deep learning, attention mechanisms, and multi-sensor fusion. Industrial case studies from the shipbuilding, automotive, and aerospace industries show how these technologies can be used practically to improve weld quality, decrease inspection efforts, and boost productivity. Nevertheless, there are still issues with sensor robustness, real-time data processing, and model generalization across various materials and joint geometries. The advancement of edge computing, adaptive learning, sensor miniaturization, and the shift to fully autonomous welding cells are the main areas of focus for future research. In order to promote autonomous decision-making, accurate seam control, and strong quality assurance in industrial welding applications, this review offers researchers and practitioners a thorough basis on which to build.</p>