Embodiments of the present disclosure disclose a method for recording sound of a video-recorded object and a mobile terminal, so as to reduce background noise in recorded sound of a video-recorded object, and improve quality of the recorded sound. The method in the embodiments of the present disclosure includes: obtaining, by a mobile terminal by means of facial recognition, position information of the video-recorded object; and performing processing on recorded sound signals, so as to enhance signal intensity of a sound signal at a position of the video-recorded object, and attenuate signal intensity of a sound signal at another position, to obtain sound at the position of the video-recorded object.