The present disclosure relates to a mobile terminal and an audio zooming method thereof capable of effectively capturing audio based on an image while capturing and reproducing video. According to the present disclosure, if an event for changing the capture location or range or capture mode of a subject occurs when capturing the subject while capturing the sound of the subject in a predetermined sound capture range, the sound of the subject may be captured by changing a sound capture angle and automatically controlling a sound capture range according to the relevant event, and if an event for changing the capture location or range or capture mode of a subject occurs, the sound of the subject may be captured by automatically controlling a sound capture range according to the event.