 This study proposes a novel multi-modality fusion assessment framework for quantifying hand motor function in post-stroke hemiplegic patients. The proposed framework consists of three components, kinematic feature extraction based on a graph convolutional network, H-A-G-C-N, surface electromygraphy, SEMG, signal processing based on a multi-layer of long short-term memory, LSTM, network, and quantitative assessment based on the multi-modality fusion. The H-A-G-C-N was used to extract kinematic features from the hand movement while the LSTM network was employed to analyze the SEMG signals. The results were then fused together to provide a comprehensive assessment of the patient's hand motor function. The proposed assessment framework was evaluated on 70 subjects who performed 28 different hand movements. The results showed a strong correlation between the proposed assessment and traditional scales, as well as a high degree of agreement between the assessments of this study and those of experienced therapists. Additionally, the sealing effect of some traditional scales could be avoided through the use of the proposed assessment. This article was authored by Chang Wangli, Han Junyang, Long Chung, and others. We are article.tv, links in the description below.