Ataxia is a clinical sign seen in several neurological diseases and has been reported as an adverse effect of anti-seizure medication such as phenobarbital (PB). Efforts to objectify canine ataxia in order to provide appropriate treatment or monitor adverse effects of drugs remain limited. Automated quantitative gait analysis can be useful for the acquisition of objective data for the evaluation and monitoring of ataxia in dogs. The aim of this prospective clinical study was to examine the gait characteristics of dogs with PB induced ataxia and compare them with healthy dogs using a computer- and treadmill-based gait analysis system. Six healthy dogs and five dogs with idiopathic epilepsy (IE) with PB-induced ataxia underwent video- and computer-assisted gait analysis during slow walking (maximum speed of 0.7 m/s) on a treadmill with four ground reaction force plates (one plate per limb). Kinetic and spatio-temporal gait parameters of dogs' locomotion were analyzed, including individually calculated coefficients of variation. Dogs with IE treated with PB showed higher variability in spatio-temporal but not in kinetic gait parameters. Double support phase of gait cycles was increased on the cost of single support and swing phases. Body weight standardized ground reaction forces in vertical, craniocaudal, and mediolateral direction were severely affected by ataxia. Compensatory mechanisms in locomotion of dogs with PB-induced ataxia included spatio-temporal and kinetic gait characteristics, most likely in order to compensate imbalance caused by limb incoordination.