An image display device with a built-in karaoke and a method for controlling the same are disclosed, wherein an audio of a broadcasted program is divided into a plurality of elements, so as to be selectively controlled when required. The method for controlling the image display device with a built-in karaoke includes receiving an output signal and demultiplexing the received signal into an accompaniment signal and a vocal signal, and selectively outputting the accompaniment signal and the vocal signal both included in the received signal, when a user selects an accompaniment mode.