Abstract: The structure-borne acoustic emission (AE) signals during pulse MIG welding on aluminum alloy were detected in real-time and analyzed to study the characteristics of AE signals corresponding to the metal droplet transfer and its influence upon the characteristic parameter of AE signals. The results showed that a mass of information about metal droplet transfer was included in the AE signals collected during PMIG welding process. It was a feasible way to characterize the behavior of metal droplet transfer by the time domain waveform and characteristic parameter of AE signals. As the frequency of metal droplet transfer is fixed, the characteristic parameters of AE signals increasee with the ineseasing of impact effect of metal droplet transferred into the molten pool. As the frequency of metal droplet transfer is unfixed, the impact effect of metal droplet to the molten pool decneases with the increasing of the droplet transfer frequency. So, the characteristic parameters of AE signals are decreased.