Pneumonia is a respiratory infection that attacks the lungs in the lung parenchyma tissue or alveoli. Pneumonia is one of the leading causes of death in children worldwide. Rational use of antibiotics can reduce the risk of antibiotic resistance. This study used observational analytics with a retrospective cohort design. The research subjects were pediatric patients with a diagnosis of pneumonia who were hospitalized at Dr. Sardjito Yogyakarta period 1 January – 31 December 2020. The rationality of the use of antibiotics was evaluated using the Gyssens method. The clinical outcome observed was the patient's condition improving or not improving after 3 to 5 days of antibiotic administration. Chi-Square test to see the relationship of antibiotic rationality to clinical outcomes and multiple logistic regression analysis to analyze the relationship of confounding variables to clinical outcomes. From 141 patients, 211 antibiotic regimens were obtained, with details of 186 empiric antibiotics and 25 definitive antibiotics. The rational use of antibiotics in empirical and definitive antibiotics was 140 regimens (75.27%) and 22 regimens (88%). Total irrational antibiotics (category I-VI) from empirical and definitive antibiotics were 24.73% and 12%, respectively. There is a significant relationship between the rationality of antibiotics with clinical outcomes both empirical and definitive antibiotics (p<0.05). Meanwhile, for confounding variables, there was no significant relationship to the clinical outcome (p>0.05). The most common bacteria were Klebsiella pneumoniae and Acinetobacter baumanii. Klebsiella pneumoniae was the most resistant to ampicillin, ampicillin sulbactam and ceftriaxone antibiotics and Acinetobacter baumanii was the most resistant ampicillin.