This paper describes an automatic welding control system developed for alternating current shielded metal arc welding (SMAW). This method could replace manual operations which require a well-trained technician. We have derived a mathematical model of the welding control system and identified the system’s parameters. The sliding surface is used as the input variable to reduce the number of fuzzy reasoning rules, in comparison with the conventional two-dimensional fuzzy logic control (FLC) algorithm. An adaptive fuzzy sliding mode controller (AFSMC) consists of an equivalent control part and a hitting control part. An adaptive law derived from a Lyapunov function is used to obtain the FLC’s parameters, and is applied to approximate the equivalent control part of the sliding mode control (SMC), so that the system states can be forced to zero. By using three-rules FLC, the control part that satisfies the hitting conditions of the SMC can force the system’s states to reach and remain on the sliding surface. Therefore, the stability of the AFSMC can be guaranteed and can be used to modulate the rate of the electrode feeding mechanism that regulates the arc current of the SMAW. The simulation and the experimental results both show that this automatic welding control system, based on the AFSMC, can perform effectively.