Previous research has shown that azotemic dogs have a lower clearance and higher drug plasma concentrations of ampicillin compared to healthy dogs. The objective of this study was to determine the pharmacokinetics of ampicillin-sulbactam after multiple intravenous doses in hospitalized azotemic and non-azotemic dogs. This prospective study included 29 client-owned dogs; 19 azotemic and 10 non-azotemic. Ampicillin-sulbactam was administered at a combined dose of 22 mg/kg intravenously every 8 h for up to 5 days. Blood samples were obtained at baseline (prior to administration of the first dose of ampicillin-sulbactam), and 1-, 4-, and 8-h post-ampicillin-sulbactam administration each day. Plasma ampicillin was measured using LC-MS and non-compartmental pharmacokinetic modeling and dose interval modeling were performed. Plasma ampicillin exposure (azotemic mean 214.5 ug/mL × h ± 110.8, non-azotemic mean 60.3 ± 35.7; p < 0.0009) and half-life (azotemic mean 3.9 h ± 2.4, non-azotemic mean 1.5 h ± 0.3; p < 0.00001) were statistically greater in azotemic dogs compared to non-azotemic dogs. Single dose interval modeling predicted that 100% of azotemic dogs would have > 50% of the dosing interval with plasma concentrations > MIC (MIC = 2) with q12 h dosing and 79% of azotemic dogs would have > 50% of the dosing interval with plasma concentrations > MIC (MIC = 8) with q12 h dosing. Comparatively, 20% of non-azotemic dogs were predicted to have > 50% of the dosing interval with plasma concentrations > MIC (MIC = 2) with q12 h dosing and 0 non-azotemic dogs would have > 50% of the dosing interval with plasma concentrations > MIC (MIC = 8) with q12 h dosing. This study demonstrated that q12-h dosing of ampicillin-sulbactam in azotemic dogs over multiple days of administration is sufficient to reach the PK-PD target (> 50% of dosing interval > MIC) against susceptible bacteria.