Objective To evaluate the validity and reliability of the social responsiveness scale short rorm （SRS-SF） and assess its accuracy  for screening autism spectrum disorder（ASD） in Chinese children. Methods Parents or caregivers of 140 clinical children who were diagnosed as having ASD based on Diagnostic and Statistical Manual of Mental Disorders （5th edition）（DSM-5） and 58 clinical children without ASD（Non-ASD） to complete the social responsiveness scale （SRS） and the parent version of strength and difficulty questionnaire （SDQ）. And 42 of them completed SRS retest after 3 months. Internal consistency reliability, testretest reliability, constructive validity and discriminate validity were used to evaluate the validity and reliability of the SRS-SF based on the SRS data. The receiver operating characteristic curve （ROC） was used to assess the accuracy of SRS-SF for screening autism spectrum disorder in Chinese children, compared with SRS. Results For the whole SRS-SF, the Cronbach's α coefficients were 0.879 and the test-retest coefficients were 0.511. The scores of the whole SRS-SF and its three factors were highly positively correlated with those of SRS（r=0.853~0.960, P<0.001）. The scores of the whole SRS-SF was moderately correlated with those of the total difficulties and prosocial behaviors of the parent version of SDQ（|r|=0.349~0.585, P<0.001）. SRS-SF had good discriminate validity among ASD and Non-ASD control groups（t=-8.367~-2.841, P<0.01）. The results in receiver operating characteristic analysis showed SRS-SF was slightly better than SRS in sensitivity（SRS-SF 0.783 vs SRS 0.704）and area-under-the-curve（SRS-SF 0.833 vs SRS 0.829）.  Conclusion These results suggest that SRS-SF has acceptable reliability and validity, and can be another useful tool for screening ASD in Chinese children.