Our objective was to evaluate the appropriateness of antibiotic prescriptions in the neonatal intensive care unit using standardized criteria and determine the effects of an antimicrobial stewardship program (ASP) on patterns of antibiotic usage. A retrospective audit of antibiotic use from July 2010 to June 2013 was conducted, focusing on prescriptions of vancomycin, cefotaxime, meropenem and linezolid for >3 calendar-days. We evaluated the appropriateness of each course of antibiotic treatment based on the Centers for Disease Control and Prevention 12-Step Guidelines to Prevent Antimicrobial Resistance (steps 4, 6 and 9). An ASP was introduced in August 2014, and the same audit criteria were applied to review antimicrobial use in the subsequent 12 months. In the pre-ASP era, 26.3%, 12.1%, 11.4% and 0% of meropenem, cefotaxime, vancomycin and linezolid courses, respectively, were inappropriate. The most common instance of inappropriate utilization included failure to use narrow-spectrum antimicrobials when appropriate. After the introduction of ASP program, 22.2%, 7.5%, 5.4% and 0% of meropenem, cefotaxime, vancomycin and linezolid courses, respectively, were inappropriate. The numbers of inappropriate antibiotic-days/1000 days of therapy with meropenem, cefotaxime and vancomycin changed from 1.89 to 1.96 rate ratio (RR), 1.04 (0.70-1.52), 3.56 to 1.73 RR, 0.49 (0.33-0.71) and 2.70 to 1.01 RR, 0.37 (0.22-0.60), respectively. In subgroup analysis, very low birth weight infants (birth weight, <1500 g) showed no improvement in the rates of inappropriate antibiotic prescriptions. In this study, we found that ASP initiatives can be applied in neonatal populations to reduce inappropriate antimicrobial prescriptions, though a different approach may be needed among very low birth weight infants.