Motion perception refers to the effect of the characteristics of the motion on the human’s brain, which was perceived and recognized by the individual. It is the result of a variety of sensory coordination activities, such as vision, kinesthesis and equilibrium sense. Researches about representational momentum have indicated that the forward displacement of the final position of a moving target is not only along its moving direction, but also along the direction of orthogonal axis, demonstrating that motion spatial orientation shows both distance and direction deviation. The present research was to explore the direction deviation under different situations, and the cognitive penetrability of motion perception by changing the meaning of the stimuli, direction of movement, and motion patterns of the stimuli. The experiment 1 adopted circle image while experiment 2 adopted the airplane image, which was designed as a simple line-drawn image. Both experiment 1 and 2 obeyed a full factorial design: 2(motion pattern: take-off, landing) × 2 (direction: rightward, leftward) × 3(moving position: same, higher, lower trajectory). Our research showed the following results. First, motion pattern significantly affected the accuracy steadily when asking to determine whether the actual trajectory of the stimulus was same with the preset one. When the object landed, people made more wrong decision. However, the effect of moving direction on accuracy depended on the meaning of stimulus. Only when the object was meaningless in experiment 1, rightward movement led to more wrong decision. When the object was the airplane image, the accuracy wasn’t different between the rightward and leftward motion. Second, moving position had an influence on participants’ judgement. We found that when actual trajectory was slightly higher than the preset one, participants were more likely to made wrong decision compared to that when the actual trajectory was lower than the preset one, p<.05, which demonstrating direction deviation in experiment 1. However, when stimulus was airplane, we found that the direction deviation occurred differently compared which was observed in experiment 1. Third, the form of direction deviation depended on the relationship between the animacy of the object and the flying scene. When the object was the airplane image, and the landing motion was performed, it was more likely for participants to make the same response when the actual trajectory was lower than the preset one. However, when take-off motion was performed for the airplane, it was more likely to make the same response when the actual trajectory was higher than the preset one. The results above indicated that participants were more likely to consider the airplane to fly lower and lower when landing movement was performed, and to fly higher and higher when take-off movement was performed. The different form of direction deviation was in accordance with the phenomena of the flight inertia. We demonstrate that when object is moving, it does deviate from its preset trajectory. However, the direction deviation isn’t always the same for take-off and landing movement, which can be influenced by the animacy of object, gravity, and motion pattern. Our results suggest that the motion perception is cognitively penetrable.