To know the effect of a postural control exercise using visual perception on the visual motor integration ability and posture change in spastic cerebral palsy children, we researched three children who had weekly classes (30 minutes per class; three times per week) for 10 weeks from June 19th to September 4th, 2011. The effect of postural control exercise using visual perception was examined by both pretest and posttest raw scores and standard scores, and maintenance tests for eye-hand coordination, drawing-over-trace, spatial relation, visual-motor velocity in subcategory of DTVP-2, and posture changes were looked at with regards to a qualitative aspect of head control motion, midline-sense, weightshift, and dissociation movement. From these results, we know that postural control exercise using visual perception positively affects the posture change of spastic cerebral palsy children. Even though normal posture and motion of cerebral palsy children with visual perception disabilities were found to have many problems, over the exercising sessions the children tried to grip objects within their sight using normal posture and motion through head control, midline-sense, weight-shift, and dissociation movement. Noticeably the children tried to look at what they intended to grip. Visual motor coordination through head control was significant for making spatial relation adjustments, such as stretching arms and crouching with trunk stability and dissociation movement through a seated posture and weight-shifting in a standing posture.