The invention relates to the technical field of image processing, in particular to a speaker recognition method, device and system based on artificial intelligence and electronic equipment. The recognition method comprises the steps of obtaining a to-be-detected image and performing face recognition processing on the to-be-detected image to obtain at least one face coordinate; identifying an audioacquisition device in the to-be-detected image to obtain first coordinate data of the audio acquisition device; calculating displacement data according to the first coordinate data and the historicalcoordinate data of the audio acquisition equipment, so as to calculate accurate coordinates of the audio acquisition equipment according to the displacement data; and calculating an inter-object distance between the accurate coordinate and the at least one face coordinate, and taking an object corresponding to the face coordinate with the minimum inter-object distance as a speaker. According to the technical scheme of the invention, after the face in the to-be-detected image and the unique audio collection device are determined, the correctness of the coordinates of the audio collection device can be judged and optimized in combination with the historical coordinate data, and the speaker recognition accuracy is improved.