Psychological measurement of special children poses a significant challenge. Due to their insufficient cognitive and self-reflective abilities, the commonly used Likert scales are ineffective in implementing psychological assessment. This chapter proposes an artificial intelligence-based framework for psychological measurement of special children, leveraging relevant models in artificial intelligence to evaluate, cluster, and predict the psychological states of these children. Case studies demonstrate the effectiveness of this artificial intelligence-based framework in psychological measurement of special children. Notably, it achieves an accuracy rate of 98% in psychological evaluation directly through externally displayed sign language movements, without the need for scales, thereby introducing novel means and approaches for psychological measurement of special children.