Rational empirical antimicrobial therapy is an important component of sepsis patient management. This study aimed to assess the rationality of empirical antimicrobial therapy in patients diagnosed with sepsis admitted in intermediate care ward of internal medicine department (RPI) of Dr. Soetomo General Hospital from January 2016 to July 2017. Medical records of 91 patients diagnosed with sepsis were collected and studied retrospectively in period from July 2017 to November 2017. 91 (85.05%) medical records from 107 sepsis patients were evaluated. Cultures and antimicrobial sensitivity tests were carried out in 21 (23.07%) patients. 14 patients yielded positive culture results, 9 of which were MDRO positive with ESBL as resistant marker. Empirical antibiotic therapies for these patients were reviewed according to Gyssens method.73 (80.2%) of 91 patients were deemed receiving appropriate empirical antibiotic therapies. Ceftriaxone IV injection as monotherapy or combination therapy were the most common empirical antibiotic therapies (82 in 91 patients, 90.1%), despite local microbiologic flora and antibiogram show most pathogens were resistant to ceftriaxone. Mortality rate in this study was high, 92.3% (84 patients died) despite rational empirical antibiotic therapies were high.This study concluded that empirical antibiotic therapies in sepsis patients according to guidelines adopted in Soetomo General Hospital, albeit deemed rational, was no longer appropriate according to local antibiogram issued by microbiological department of Soetomo General Hospital.