Objective To explore the association of coefficient of glycemic variation (CV) and score for neonatal acute physiology with perinatal extension-Ⅱ (SNAPPE-Ⅱ) with prognosis in critically ill neonates. Methods The clinical data of critically ill neonates admitted to the neonatal intensive care unit (NICU) from August 2018 to September 2020 were retrospectively analyzed. According to the prognosis, they were divided into death group and survival group, and the difference of clinical features between the two groups was compared. Binary logistic regression model was used to analyze the risk factors of poor prognosis in critically ill neonates. Receiver operating characteristic (ROC) curve analysis was used to evaluate the diagnostic value of CV and SNAPPE-Ⅱ in poor prognosis of critically ill neonates. Results A total of 385 critically ill newborns were included, including 228 boys and 157 girls. The median age of admission was 1.0 (1.0-24.0) hours, the mean gestational age was (35.3±4.1) weeks, and the mean weight was (2456.6±860.3) grams. There were 85 hypoglycemic and 115 hyperglycemic patients. There were 80 patients in the death group, and the median time from NICU admission to death were 5.0 (2.0-10.3) days. Binary logistic regression analysis showed that CV, SNAPPE-Ⅱ, gestational age, weight, the main diagnosis of pulmonary hemorrhage and neonatal sepsis were correlated with death of critically ill neonates (P<0.05). The area under curve (AUC) of CV predicting death of critically ill neonates was 0.72, and the AUC of SNAPPE-Ⅱwas 0.85. When CV≥0.31, the sensitivity and specificity for the diagnosis of critical neonatal death were 0.75 and 0.59. When SNAPPE-Ⅱ≥18.50, the sensitivity and specificity for the diagnosis of critical neonatal death were 0.68 and 0.93. Conclusions Both CV and SNAPPE-Ⅱ have certain predictive values for the poor prognosis of critically ill neonates.