Objective: To study the case mix mode and cost payment standard of patients with premature infants，and to provide a reference for formulating hospitalization expenses groups and improving medical efficiency. Methods: A total of 3，061 cases of premature infants in a tertiary maternity and children's hospital from 2019 to 2021 were collected. Multiple linear stepwise regression analysis was used to determine the main influencing factors of hospitalization expenses. The CHAID method in decision tree model was used to construct the hospitalization expense grouping plan and payment standard for premature infants. Results: The number of days in hospital，number of comorbidities or complications，neonatal birth weight，presence or absence of operations，length of time on ventilator，and transfer status were used as classification nodes to influence the hospitalization cost of preterm cases，and a total of 7 DRG combinations and corresponding hospitalization cost criteria were formed using a decision tree model to group 3，061 preterm patients. Conclusion: Case-mix diagnosis related groups of premature infants could objectively reflect the level of medical resource consumption，and provide a reference for medical institutions in other regions to verify the grouping and optimize the grouping plan.