Abstract: Because of the limitations in traditional acoustic emission positioning algorithms of rock mechanics experiments,we introduce the gray wolf algorithm(GWO)into the acoustic emission positioning research. This algorithm imitates the leadership and hunting mechanism of wolves through simulating head wolf leadership. On the basis of the original gray wolf algorithm,aiming at the defect of its poor local search ability,we propose an improved gray wolf algorithm(BGWO)based on the population memory elimination system. The BGWO algorithm modifies the method of decreasing the convergence factor and eliminates the worst individual. We compare and analyze the performance of five positioning algorithms. They are BGWO,GWO,Gravity search algorithm(GSA),Geiger algorithm and LS algorithm in the experimental of uniaxial compression acoustic emission of prefabricated fractured rock specimen. We find out that the BGWO algorithm is better than other algorithms in terms of acoustic emission positioning search efficiency,search accuracy,stability and simulation of experimental results.