Neonatal sepsis is a clinical syndrome of systemic disease, accompanied by bacteremia that occurs during infants in the first month of life. A late diagnosis might increase mortality. The presence of bacteria growth in blood cultures is a definitive diagnosis. Unfortunately, culture results are usually obtained of a long time. The study aimed to analyze sensitivity and specificity of the manual I/T ratio, automatic I/T ratio, leukocyte count and procalcitonin (PCT) to diagnose neonatal sepsis. This study used a cross-sectional design, from the NICU room in Sanglah General Hospital, Denpasar. There were 59 patients who met the study criteria. Along with blood culture as the gold standard in determining diagnosis of sepsis and with I/T ratio cut-off of 0.2, the sensitivity of manual I/T ratio was 69.2%, specificity 83.9%, PPV 63.9%, NPV 87% and likelihood ratio was 3.06. While the sensitivity of automatic I/T ratio was 47.6%, specificity 85.8%, PPV 55.1%, NPV 81.4% and likelihood ratio was 2.25. Based on the normal range of leukocyte count (9.1 - 34 x 103/μl), sensitivity of leukocyte count was 59%, specificity 71.5%, PPV 46.7%, NPV 80.9% and likelihood ratio was 1.59. With PCT cut-off 0.5 ng/mL, the obtained sensitivity of PCT was 64.3%, specificity 85.8%, PPV 64.3%, NPV 85.8% and likelihood ratio was 3.13.