Eye tracking of preferential attention to biological motion is an important approach to Autism Spectrum Disorder (ASD) research and early diagnosis. However, the current method has all time-series information averaged as statistic information, which losing information and features for classification in high-dimensional space. Novel biological motion stimuli with well-designed time-series information could make Insistence on Sameness (IoS) of ASD as an efficient feature for ASD screening. Display area zoning and encoding for eye-tracking data processing improve the spatial resolution of the eye tracking information processing than averaging by side, making it possible to get more information and features for classification from ASD eye-tracking data.