This paper proposes the viewpoint change method that uses a 3-D mouse for stereoscopic vision modeling based on the movement of objects. Our previous modeler offered only a stereoscopic perspective view. However, it was diffcult for users to precisely create and place objects with this modeler. Based on usage studies, we created an improved modeler that adds three orthographic views to the previous modeler. Examinations showed that 3-D mouse operations do not always directly correspond to cursor movements on the screen due to camera viewpoint change. We propose a new viewpoint change method that is based on the object-move method, which makes 3-D mouse movements more intuitive. We also describe tests that confirm the effectiveness of the proposed method. The results show that the proposed method is especially effective for novices.