This study investigated the formation process of “motor imagery” (including “visual imagery” and “kinesthetic imagery”) concerning the handstand, and its relation to the metacognitive perspective. The results showed that the number of metacognitive perspectives changed from unidirectional, single perspective to multidirectional, multiple perspectives, and then reverted back to unidirectional, single perspective. The position of the metacognitive perspective shifted from a relatively narrow focus close to the body to a slightly broader focus at a greater distance. Additionally, the movement pattern transitioned from fixed points to point-to-point movement, then to linear movement, and finally returned to a fixed state. Furthermore, the speed of the movement of the metacognitive perspective increased from relatively slow point-to-point movement to instantaneous movement, before reverting to a slower speed for point-to-point movement. Accompanying these changes in the position and movement of the metacognitive perspective, the range and spatial positioning of the vividly recalled “visual imagery” expanded from limited body parts to a broader area and eventually to the whole body. Moreover, “kinesthetic imagery” became controllable from the areas where the “visual imagery” became clearer.