The Acoustic Emission testing application on Crane steel structure is greatly limited, because of the factors such as the random transmission path of sound wave, the influence of dynamic load noise and the discontinuity of structure. In this paper, the feasibility of Acoustic Emission detection for crane structure damage is discussed from two aspects: mechanism test and field detection. Based on the analysis of the whole test process, when lifting heavy objects and braking, the metal structure is subjected to the instantaneous gravity load of tens of tons, so that a large number of AE signals are collected by various monitoring sensors. It mainly includes active defect damage, mechanical vibration, structural friction and electrical noise, etc. At this time, the effective defect expansion signal is compared with the noise signal. However, in the load maintenance phase, the noise signal disappears or drops to a very low level. At this point, if a sensor still collects a strong active AE signal, it is highly likely that there is a damage source in this area. Compared with the traditional mechanical properties of materials, AE characteristic parameters, such as amplitude, ringing count and energy count, can reflect the microscopic damage changes of materials under load in a more detailed way. By summing up the distribution range of AE parameters corresponding to different damage mechanisms and typical signal characteristics, such as the double peaks phenomenon in the material yield stage, it can provide a scientific foundation for the application of AET in the metal structure damage of large lifting machinery.