Symptoms and signs of early infection and sepsis in premature infants are varied; there are no clear criteria for the diagnosis of such conditions in this cohort. The aim of our study was to analyze and identify the risk factors for early infections in premature infants and to develop on their basis the clinical prognostic model with high diagnostic performance. A retrospective cohort study, which included 152 premature infants, was conducted; 121 of them had the signs of early infections, 31 - had no signs of infection. 52 candidates of prognostic variables were considered. According to the results of multiple stepwise logistic regression analysis, the predictive model has been developed. It includes gestational age, visual changes of the placenta, Apgar score at the 1st minute, the level of monocytes more than 6.5%, the history of abortions and premature rupture of membranes. The diagnostic characteristics of the developed model are high: sensitivity - 82.2%, specificity - 93.55%, positive predictive value - 97.98%, negative predictive value - 58%. Therefore, when interpreting the figures of biomarkers, the decision as to the prescription of antibiotics should be based on the presence of maternal risk factors and clinical symptoms of the infection in the newborn.