This study is conducted to carry out a straightforward way appropriate for quality monitoring and stability of arc stud welding process, followed by a number of procedures to control the quality of welded samples, namely torque destructive testing and visual inspection context.  Those procedures were being performed to support the monitoring system and verify its validity. Thus, continuous on-line monitoring guarantees earlier discovering stud welding defects and avoiding weld repeatability. On-line welding electronic monitoring system is for non destructive determining if a just completed weld is satisfactory or unsatisfactory, depending on welding current peak value detected by the system. Also, it has been observed significant harmonize which is mutually linking the monitored current peak values and quality control measures. So this concept is accordingly contributed in the process of supporting the fundamental objective of this research. On the other hand, two feed-forward neural networks have been developed for monitoring and control arc stud welding quality. First network predicts two output quality parameters (current peak value) and (torque testing value at failure). Second, predicts one output quality parameter (visual inspection). Networks have been trained to a set of data, which made them ready to receive new information for subsequent quality parameters prediction. Both networks showed up good response and acceptable results.