Objective <p>Neuroserpin, a serine protease inhibitor, is recognized for its anti-inflammatory and neuroprotective properties. Given the central role of inflammation and neurological involvement in the pathophysiology of preeclampsia, this study aimed to assess maternal serum neuroserpin levels in preeclamptic pregnancies and explore their association with disease severity.</p> Design <p>Prospective case-control study.</p> Setting <p>A tertiary referral center in Ankara, Türkiye.</p> Population <p>Singleton pregnant women with a diagnosis of preeclampsia (<i>n</i> = 44) and gestational age-matched normotensive pregnant women as controls (<i>n</i> = 44).</p> Methods <p>Participants were assigned to preeclampsia and control groups. Serum neuroserpin levels were quantified using enzyme-linked immunosorbent assay (ELISA). Subgroup analysis was conducted based on the clinical severity of preeclampsia. Statistical analysis included group comparisons, receiver operating characteristic (ROC) curve analysis, and correlation testing.</p> Main Outcome Measures <p>The primary outcome was maternal serum neuroserpin level. Secondary outcomes included obstetric and neonatal parameters such as gestational age at delivery, delivery mode, NICU admission, and Apgar scores.</p> Results <p>Serum neuroserpin levels were significantly reduced in the preeclampsia group compared to controls (<i>p</i> = 0.018). Within the preeclampsia cohort, patients with severe disease exhibited even lower neuroserpin concentrations than those with mild preeclampsia. ROC curve analysis determined a neuroserpin cutoff value of ≤ 22.95 ng/mL for identifying preeclampsia (AUC: 0.647, <i>p</i> = 0.013) and ≤ 14.7 ng/mL for severe preeclampsia (AUC: 0.740, <i>p</i> = 0.007).</p> Conclusion <p>Reduced maternal serum neuroserpin levels are associated with both the diagnosis and severity of preeclampsia. These findings highlight the potential role of neuroserpin in the disease’s inflammatory mechanisms and support its utility as a candidate biomarker in clinical prediction models.</p> Clinical trial registration <p>Not applicable</p>