Abstract: Objective To compare the ability of Neonatal Critical Illnesss Score (NCIS),Score for Neonatal Acute Physiology-Ⅱ (SNAP-Ⅱ) and Score for Neonatal Acute Physiology-Ⅱ and Perinatal Extension (SNAPPE-Ⅱ) in predicting the mortality of neonates with neonatal respiratory distress syndrome (NRDS). Methods From January 2017 to June 2018,237 neonates diagnosed with NRDS and admitted to the NICU of the Southern Hospital were selected and they were divided into survival group and death group according to their outcome at the time of discharge.All neonates were evaluated with NCIS,SNAP-Ⅱ and SNAPPE-Ⅱ,and the ROC curves of the three scoring systems for predicting the mortality were described,then their AUC values were compared. Results The AUC values of NCIS,SNAP-Ⅱ and SNAPPE-Ⅱ were 0.859 (95% CI:0.808-0.901),0.767 (95% CI:0.708-0.819) and 0.798 (95% CI:0.741-0.847),without statistically significant difference (P>0.05).The cut-offs of SNAP-Ⅱ and SNAPPE-Ⅱ were 9.5 (sensitivity 72.22%,specificity 77.63%) and 19.5 (sensitivity 83.33%,specificity 66.21%).The mortality of non-critical group,critical group and extremely critical group by NCIS were 0,8.43% and 66.67%,with statistically significant differences (P<0.05).The mortality of critical neonates met the NCIS individual index or score index was 0 and 3.4%,which was significantly lower than that of neonates both met individual index and score index by 14.8% (P<0.05). Conclusion The ability of NCIS,SNAP-Ⅱ and SNAPPE-Ⅱ to predict the mortality of the neonates with NRDS is consistent.NCIS is appropriate for the classification of disease severity,and the combination of individual index and score index is more significant in predicting severity.