A method and apparatus to process a user-defined audio user interface (UI), and an audio device using the same. The method includes designating a specific portion of an audio signal being reproduced, using a specific input member, capturing audio data of the designated specific portion, mapping the captured audio data of the designated specific portion to information on the specific input member, and constructing an audio UI database using the audio data of the designated specific portion mapped to the information on the specific input member. Accordingly, by constructing an audio UI database by selecting sounds preferred by a user, demands of various users, which cannot be provided by existing MP3 players, can be satisfied.