An ear speaker device includes an electroacoustic transducer including a housing mounted at a predetermined position of listener's head, a speaker unit mounted on one surface in the housing, and positioned away from an entrance of the listener's external acoustic meatus for a predetermined distance when the housing is mounted on the listener's head, and a tubular duct extended so as to allow a sound generated by the housing in the inner space thereof to reach the vicinity of the entrance of the listener's external acoustic meatus, and a hole for emitting sound of the duct is oriented to the opposite direction with respect to the entrance of the external acoustic meatus; and a mounting part used for mounting the electroacoustic transducer on the listener's head in a manner that the predetermined distance is provided between the speaker unit and the entrance of the listener's external acoustic meatus.