Objective：To investigate the predictive value of oxygenation parameters and ventilator parameters on the clinical outcome of acute respiratory distress syndrome(ARDS)at different times，and to screen early prognostic indicators. Methods：The clinical data of 65 children with ARDS admitted to the intensive care unit of Children’s Hospital of Nanjing Medical University from January 2017 to June 2021 were collected. According to the survival status after 28 days of treatment，they were divided into survival group(n=37)and death group(n=28). Logistic regression analysis of clinical data was to identify risk factors for death in ARDS patients. Receiver operating characteristic(ROC)curve was used to evaluate the predictive value of oxygenation parameters and ventilator parameters on mortality risk in children with ARDS at different times. Results：The median age of the 65 patients was 2.1 years old. Pneumonia(60%)and sepsis(24.6%)were the main causes of ARDS. Analysis of the underlying disease status of patients showed that 60% of patients had underlying diseases，including neurological diseases(21.5%)，hematological diseases(16.9%)，and chronic lung diseases(6.2%)and immunodeficiency diseases(7.7%). Logistic regression analysis showed that PaO₂/FiO₂，SpO₂/FiO₂， OI，OSI，MAP and PEEP were independent risk factors for ARDS on days 0，1 and 3 of mechanical ventilation(P ＜ 0.05). The above indicators showed better predictive value on the 3rd day of mechanical ventilation(AUC：0.806~0.961). Conclusion：PaO₂/FiO₂，SpO₂/ FiO₂，OI，OSI，MAP and PEEP are independent risk factors for death in children with ARDS，and each index has better predictive value on the third day of mechanical ventilation.