In this study, we introduce an asymmetry calculation model for evaluating the effectiveness of facial paralysis diagnosis and treatment, implemented through a Python and OpenCV algorithm. The algorithm is based on actual images from cases of facial paralysis treatments, and numerical experiments for image comparison and subsequent detailed statistical analysis are conducted. The research findings indicate that the calculated numerical indices provided by this model can relatively accurately assess the effectiveness of facial paralysis diagnosis and treatment on a graded scale. Consequently, in this study, we propose to combine overall facial and specific facial region motion disparity features for a comprehensive facial paralysis grading evaluation. This innovative approach provides a robust tool for the accurate assessment of facial paralysis treatment outcomes, offering significant support for clinical practices and treatment optimization.