The present invention relates to a crosswalk voice guide aid device, comprising: a pedestrian sensing unit, installed in a pedestrian signal light post or a separate post in a crosswalk, for sensing a pedestrian's entry into a sensing zone in any of the cases where a pedestrian signal light is a stop signal, a walk signal, or a flashing walk signal, or a signal light does not exist; an entry determination reference unit formed on a pedestrian of an image sensed by the pedestrian sensing unit; a pedestrian entry determination unit for determining whether the pedestrian has or has not entered the sensing zone, through the entry determination reference unit, and for transmitting a corresponding signal to a control unit when it is determined that the pedestrian has entered the sensing zone; a control unit for causing a guide voice corresponding to the sensing zone entry to be outputted through a voice output unit, when the signal received from the pedestrian entry determination unit is a signal indicating that the pedestrian has entered the sensing zone; and a voice output unit for outputting a guide voice under the control of the control unit, wherein the entry determination reference unit is formed at a position corresponding to a pedestrian's leg region, which is a position at a predetermined height from the ground, of the pedestrian of the image sensed by the pedestrian sensing unit, or at a position on the ground where the pedestrian of the image sensed by the pedestrian sensing unit stands.