Intravenous benzodiazepines are utilized as first-line drugs to treat prolonged epileptic seizures in dogs and alternative routes of administration are required when venous access is limited. This study compared the pharmacokinetics of midazolam after intravenous (IV), intramuscular (IM), and rectal (PR) administration. Six healthy dogs were administered 0.2 mg/kg midazolam IV, IM, or PR in a randomized, 3-way crossover design with a 3-day washout between study periods. Blood samples were collected at baseline and at predetermined intervals until 480 min after administration. Plasma midazolam concentrations were measured by high-pressure liquid chromatography with UV detection. Rectal administration resulted in erratic systemic availability with undetectable to low plasma concentrations. Arithmetic mean values ± SD for midazolam peak plasma concentrations were 0.86 ± 0.36 μg/mL (C0) and 0.20 ± 0.06 μg/mL (Cmax), following IV and IM administration, respectively. Time to peak concentration (Tmax) after IM administration was 7.8 ± 2.4 min with a bioavailability of 50 ± 16%. Findings suggest that IM midazolam might be useful in treating seizures in dogs when venous access is unavailable, but higher doses may be needed to account for intermediate bioavailability. Rectal administration is likely of limited efficacy for treating seizures in dogs.