Introduction: To identify the risks and outcomes for multidrug-resistant Gram-negative bacilli (MDRGNB) sepsis in neonates. Methodology: This was a retrospective case-case-control study between 1991 and 2016. The control group was selected from the same source records of all neonates with clinical or suspected sepsis but not culture-proven. Results: The numbers of patients in the MDRGNB sepsis, non-MDRGNB sepsis, and control groups were 157, 88, and 218, respectively. MDRGNB sepsis was significantly associated with outborn infants [adjusted odds ratio (aOR) 2.08; p = 0.003] and infants who had a neurologic sequela (aOR 11.58; p = 0.04), lower gestational age (p = 0.03) or previous aminoglycoside use (aOR 2.43; p < 0.001) compared with the control group. Non-MDRGNB sepsis was associated with outborn infants (aOR 2.63; p < 0.001), and infants who had neurologic sequelae (aOR 48.25; p = 0.001) and previous cephalosporin use (aOR 6.28; p < 0.001) or cefoperazone plus sulbactam use (aOR 6.48; p = 0.02) compared with the control group. Case fatality (OR 3.63; p < 0.001) and septic shock (OR 12.81; p < 0.001) rates, length of stay (p < 0.001), and daily hospital costs (p = 0.01) were higher in the MDRGNB sepsis group than in the control group. Conclusions: Smaller preterm neonate with previous aminoglycoside use had a higher MDRGNB than non-MDRGNB sepsis compared with the control group. Intervention to reduce MDRGNB sepsis in the NICU is cost-effective.