Introduction: Neonatal sepsis is a major cause of morbidity and mortality among newborns. Sepsis accounts for 30-50% of all neonatal deaths in developing countries. C-reactive protein (CRP) has been used to predict neonatal sepsis with high sensitivity and specificity. Currently, blood culture is the gold standard for neonatal sepsis diagnosis, but it takes at least 48 hours to 7 days and not all cases have positive culture in neonatal infection. Objective: To assess the performance of C-reactive protein for predicting neonatal sepsis. Methodology: A prospective and retrospective descriptive study in neonates aged up to 28 days old presenting with symptoms and signs of sepsis and maternal high-risk factor of neonatal sepsis at Pediatric Intensive Care Unit of Mahosot Hospital. C-reactive protein rapid test was considered positive if the test level was ≥6mg/L. Blood culture was used as reference test. Result: Among 725 neonates recruited in the study, 3.9% blood cultures were positive. Among these, 21.4% were Staphylococcus aureus, 14.3% Escherichia coli and 10.7% Klebsiella pneumoniae. Sensitivity, Specificity, Positive and Negative predictive values of CRP in predicting neonatal sepsis were 71.42%, 70.15%, 8.77% and 98.3% respectively. Conclusion C-reactive protein rapid test could not predict neonatal sepsis very well, but might enable to rule out non-sepsis patients because of its high negative predictive value. This is clinically useful for treatment follow-up and antibiotic use.