Objective To investigate the early predictive value of lymphocyte-to-monocyte ratio (LMR), neutrophil-to-lymphocyte ratio (NLR), and platelet-to-lymphocyte ratio (PLR) within 24 hours after birth for retinopathy of prematurity (ROP). Methods A total of 78 preterm infants admitted to Zhengzhou First People's Hospital from January 2020 to January 2024 were retrospectively enrolled. The results of the first complete blood count within 24 hours after birth were collected, and LMR, NLR, and PLR were calculated based on the raw blood cell counts. According to the results of the first fundus screening, the infants were divided into an ROP group and a non-ROP group. Using ROP detection as the dependent variable, nested binary logistic regression models were constructed: Model 1 (clinical baseline model) included gestational age and birth weight based on clinical importance, and Model 2 (extended model) added LMR, NLR, and PLR to Model 1. Receiver operating characteristic (ROC) curves were plotted, and the areas under the curve (AUC) were calculated. The DeLong test was used to compare the differences in AUC between models, and 1000 bootstrap resamples were used for internal validation. Results Among the 78 preterm infants, 31 cases of ROP were detected, with a detection rate of 39.74% (31/78). Within 24 hours after birth, the lymphocyte count and LMR in the ROP group were higher than those in the non-ROP group, while the neutrophil count, monocyte count, NLR, and PLR were lower than those in the non-ROP group (all P<0.05). The results of the nested binary logistic regression analysis showed that in Model 1, gestational age was associated with ROP detection (OR=0.646, 95% CI: 0.504-0.829, P=0.001). After adjusting for gestational age and birth weight in Model 2, only NLR was independently associated with ROP detection (OR=0.548, 95% CI: 0.311-0.964, P=0.037). Compared with Model 1, the overall fit of Model 2 improved (P=0.009). The ROC analysis results showed that the AUCs of LMR, NLR, PLR, Model 1, and Model 2 were 0.714, 0.759, 0.667, 0.787, and 0.841, respectively. The AUC of Model 2 was higher than that of Model 1, but the DeLong test showed no statistically significant difference (P=0.147); after bootstrap correction, the AUC of Model 2 was 0.805. Conclusion LMR, NLR, and PLR within 24 hours after birth were associated with ROP detection, and NLR remained independently associated with ROP after adjusting for gestational age and birth weight. The extended model including gestational age, birth weight, and LMR, NLR and PLR within 24 hours after birth had moderate predictive performance, but with limited incremental predictive value.