Journal of Intelligent Computing Zhang Limei 16 2 2025 https://doi.org/10.6025/jic/2025/16/2/47-59 https://www.dline.info/jic/fulltext/v16n2/jicv16n2_1.pdf This article studies the rehabilitation capture analysis method for limb motion training based on human motion capture. This method aims to improve the effectiveness of exercise training and the quality of rehabilitation by capturing the motion characteristics of the human body and conducting training and rehabilitation analysis on limb movements. On this basis, this article proposes a limb motion rehabilitation analysis method based on fuzzy decision trees. This method utilizes fuzzy set theory and decision tree algorithm to evaluate and predict the rehabilitation effect of limb movements. Specifically, a series of fuzzy rules are first set based on experience, and then the membership degrees of each rule are calculated based on the collected motion data. Finally, a decision tree model reflecting the rehabilitation effect is constructed using the decision tree algorithm. The experimental results indicate that the rehabilitation capture analysis method for limb motion training based on human motion capture can accurately evaluate and predict the rehabilitation effect of limbs, providing strong support for exercise training and rehabilitation treatment.