Objective:To investigate the reliability, validity and sensitivity of Chinese scale of clinical neurological deficit of stroke patients (China Stroke Scale, CSS), so as to assess its clinical application value.Methods: A total of 126 consecutive inpatients with acute stroke onset were included in our study and they were scored by CSS and the United States National Institutes of Health Stroke Scale (NIHSS) score separately; the reliability, validity, and sensitivity of CSS were evaluated. Reliability was evaluated by correlation coefficient r and Cronbach’s α coefficient; construct validity was analyzed by factor analysis method of appraisal; criterion validity was analyzed by the correlation coefficient analysis with NIHSS scale as the criterion. Sensitivity in various fields was assessed through standardization of effect (SES).Results: Totally 123 valid questionnaires were collected. CSS showed high intrarater reliability, interrater reliability (0.911-1.000) and good internal consistency, with the Cronbach’s α>0.8. There was concurrent validity between CSS and NIHSS (r=0.86). The prognosis prediction accuracy of CSS was 92.4%, slightly lower than that of NIHSS (94.1%). Logistic regression showed that CSS’s “gaze function” and “facial paralysis” were not included in the prediction equation. The facial paralysis had a SES of 0.38, all others had a SES higher than 0.5. Most fields showed a good sensitivity.Conclusion: CSS shows an acceptable reliability, validity and sensitivity in patients with stroke, but the predicative validity of CSS is inferior to that of NIHSS, which needs be further revised.