Abstract: For the CO₂ welding of dip transfer,the electrical signals of the welding process are not only random,but also the carrier of abundant information of various physical phenomena of arc.For the random signal using the statistical analyzing method is easy to transfer the signal into the compact and accessible format,which can reflect the situation of the whole welding process.Mean while the characteristic information interrelated with the welding result can be extracted out and showed as exact figures and tables,basing on which it is hopeful to objectively evaluate the dynamic characteristic of CO₂ arc welding power source.Therefore,in this paper,the principle of the statistical analyzing method of arc welding is discussed.The experimental research of the dynamic characteristic of CO₂ arc welding power source is carried out.On these bases,an adopting statistical analysis extraction method of the characteristic information for the evaluating dynamic characteristic is studied.The statistical characteristic information and vector for the evaluating dynamic characteristic are extracted at last.The studying results show that when the dynamic characteristic is changing,the markedness of the characteristic information in the statistical characteristic vector also changes according to a definite rule.This result establishes a base for the statistical analyzing evaluation method which can be used to evaluate the dynamic characteristic of the arc welding power source according to the change of the statistical characteristic informations.