Background <p>Bronchopulmonary dysplasia (BPD) is the most common chronic lung disease in preterm infants. Although inflammation is broadly recognized as a key contributor to BPD, the precise roles and underlying mechanisms of immune cells in BPD pathogenesis remain incompletely understood. Moreover, the absence of reliable early diagnostic tools impedes the prediction of disease progression and the implementation of timely interventions.</p> Methods <p>We first analyzed public blood transcriptomic data from preterm infants (&lt;32 weeks gestational age) to explore immune alterations in BPD. Subsequently, we isolated neutrophils from infants at postnatal day 14 for transcriptomic and flow cytometric analyses. Longitudinal complete blood count (CBC) data were retrospectively collected at birth, and on days 2, 7, 14, 21, and 28, to validate these findings; machine learning models were subsequently developed to predict disease severity, with performance evaluated using the area under the receiver operating characteristic curve (AUC). Finally, single-cell transcriptome analysis was performed on public lung data from hyperoxia-exposed neonatal mice, followed by experimental validation.</p> Results <p>Blood transcriptomic profiling revealed increased neutrophils and downregulated T cell response pathways in BPD patients. Neutrophils from BPD infants exhibited enhanced inflammatory signaling and downregulated T-cell activation pathways compared to preterm controls. Flow cytometry confirmed an increased presence of polymorphonuclear myeloid-derived suppressor cell-like (PMN-MDSC-like) neutrophils in BPD, further supporting the presence of immunosuppression-associated features. Longitudinal CBC data (n = 381 infants) confirmed increased neutrophils and decreased lymphocyte levels in BPD, which correlated with disease severity. Based on this, we developed machine learning models, which exhibited promising internal-test performance in predicting moderate-to-severe BPD (AUC = 0.9328). In BPD mouse lungs, single-cell analysis showed increased neutrophils with elevated immunosuppressive signatures, including <i>Cd274</i> and <i>Lgals3</i>, and decreased CD4⁺ and CD8⁺ T cells. Flow cytometry validated these findings; neutrophil depletion in vivo attenuated alveolar simplification and led to an increase in CD4⁺ and CD8⁺ T cells in hyperoxia-exposed neonatal mice.</p> Conclusion <p>Our study identifies neutrophil-associated immune signatures in human BPD and supports a contributory role of neutrophils in hyperoxia-induced alveolar simplification in mice. Furthermore, we demonstrate the potential of longitudinal CBC-based biomarkers for disease severity stratification.</p>