Industrial welding robots or mechanized equipment for high volume manufacturing have been replacing the manual welding processes, increasing the productivity, and minimizing the harsh work environment and extreme physical demands. However, there are several problems that need to be considered for semi-autonomous robotic welding to produce welds of desired quality. Some of the most common problems are the need to compensate for inaccuracies of fixtures for the workpiece to be welded, workpiece-to-workpiece dimensional variations, imperfect edge preparation, and in-process thermal distortions. Such problems can be solved using sensory feedback in real time along with suitable adaptive or intelligent control systems so that proper joint edge detection, joint seam tracking, and weld penetration control can be achieved accounting for variation in joint location and geometry. This paper, based on a review of wide literature, presents some of the essential considerations in the selection of appropriate sensors, their need in robotic welding, the classification of weld sensors commonly used, and their applications in robotic arc welding. This research familiarizes the fabrication metal industries with the sensors, their utility, and the selection criteria for their implementation in robotic welding.