A method for determining an utterance direction includes: executing a first calculation process that includes determining an utterance direction region based on a phase difference of each given frequency between a first frequency signal and a second frequency signal inputted to a first and second voice input section respectively, and performing a process of calculating a first phase difference deviation degree indicating the degree by which the phase difference of each given frequency is deviated from the utterance direction region; executing a second calculation process for calculating a second phase difference deviation degree from the first phase difference deviation degree of the plurality of frames; and executing an utterance direction determination process for determining that a user is uttering a voice to the first voice input section and the second voice input section when the second phase difference deviation degree is equal to or smaller than a first threshold value.