Necrotizing enterocolitis (NEC) is a life-threatening disease that affects premature infants. However, the role of inflammatory biomarkers in identifying surgical/death NEC without pneumoperitoneum remains elusive. We aimed to verify the value of platelet-to-lymphocyte ratio (PLR) and the combination of white blood cell (WBC), absolute neutrophil count (ANC), absolute lymphocyte count (ALC), neutrophil lymphocyte ratio (NLR), PLR, C reactive protein (CRP) and procalcitonin (PCT) in predicting the severity of NEC, and to construct a model to differ surgically NEC from non-surgically NEC. A retrospective analysis was performed on 191 premature infants with NEC. Based on the inclusion and exclusion criteria, 90 infants with Stage II and IIIA NEC were enrolled in this study, including surgical/death NEC (n = 38) and medical NEC (n = 52). The values of inflammatory biomarkers were collected within 24h of onset. The univariate analysis revealed that the values of WBC (p = 0.040), ANC (p = 0.048), PLR (p = 0.009), CRP (p = 0.016) and PCT (p < 0.01) in surgical/death NEC cohort were significantly higher than medical NEC cohort. Binary multivariate logistic regression analysis indicates that ANC, PLR, CRP, and PCT are capable of distinguishing infants with surgical/death NEC, and the AUC of the regression equation was 0.79 (95% CI 0.64-0.89; sensitivity 0.63; specificity 0.88), suggesting the equation has a good discrimination. Elevated PLR is associated with severe inflammation in surgical/death NEC patients. The prediction modelling of combination of ANC, PLR, CRP and PCT can differentiate surgical/death NEC from infants with medical NEC, which may improve risk awareness and facilitate effective communication between nurses and clinicians. However, multicentre research is needed to verify these findings for better clinical management of NEC.