Introduction: Neonatal sepsis is a major cause of morbidity/mortality and the definitive diagnosis is an isolation of the pathogen from blood culture which might take 2-7 days. A test for early confirmation of infection is therefore required. C-reactive protein (CRP) is therefore suggested as an early screening tool in the diagnosis of neonatal sepsis. Aim: To determine the usefulness of CRP in the early diagnosis of neonatal sepsis among neonates. Study design: This was a prospective longitudinal study. Place of the study: Department of Paediatrics, Obafemi Awolowo University Teaching Hospital Complex Ile-Ife. Methods: Consecutive neonates were recruited. Blood culture was done and CRP was done at contact and 24 hours. Data were analysed and P = .05 was considered significant. Results: A total of 180 neonates comprising 106 (58.9%) males with a male to female ratio of 1.4:1 were studied. Thirty-two (17.8%) of the neonates had culture-proven sepsis with a prevalence of 10.1% among the inborn and 23.8% among the out-born with statistically significant difference (χ2 = 5.638, P = .018). The means of initial and repeat CRP for subjects with culture-proven sepsis were 41.4 (23.6) mg/l and 10.6 (4.3) mg/l respectively while subjects without sepsis were 9.2 (11.3) mg/l and 6.1 (2.6) mg/l respectively (P < .001). The CRP has a sensitivity of 93.8%, specificity 91.9%, negative predictive value 98.6%, and positive predictive value of 71.4%. The area under the curve for the receiver operator characteristic curve for subjects with CRP ≥ 10mg/L and positive culture was 0.909 (P < .001). Conclusion: The CRP has a high sensitivity, specificity, and negative predictive value and can therefore be used to screen neonates with sepsis.