An image observation apparatus comprising an image display unit for displaying an observation image and a supporting mechanism for supporting the image display unit movably and adjustably in the directions of three, substantially perpendicular axes. Since the supporting mechanism of this image observation apparatus can move and adjust the image display unit in three orthogonal axial directions, an operator need not select a plurality of articulates for operating them when he/she wants to move the image display unit to a desired axial direction. Further, one of the three axes can be a rotational axis for turnably supporting the image display unit around the operator. According to this configuration, the image display unit can be independently moved and adjusted in three axial directions, while the observation angle of the observer with respect to the display face of the image display unit is kept constant.