Abstract: A novel method,independent component analysis(ICA),was introduced to gas metal arc welding(GMAW) process on-line monitoring.ICA was applied to arc signals to remove the correlation between welding current and arc voltage to get independent component(IC).Experiments were carried out to study the feasibility of ICA for on-line monitoring of GMAW process.It is found that IC puts up an abnormity when there is a step disturbance in the welding process.Experimental results showed that IC could be used as a state variable representing the process variation.By using statistical process control for the obtained IC,the burning-through defect can be isolated from the normal situation.The comparison between ICA and principal component analysis was also done for the process.