Disclosed are a method of controlling a source device in a sink device and apparatus therefor. The present invention includes the steps of outputting the video data received from the source device to a display unit, recognizing a text from a screen corresponding to the video data outputted to the display unit, saving the recognized text to a memory to correspond to location information on the screen corresponding to the video data, receiving a voice signal, converting the received voice signal into a text, and if a specific text including at least one portion of the converted text exists in the recognized text, sending a message corresponding to a user input of selecting the specific text from the screen corresponding to the video data to the source device.