We studied the time necessary to obtain reliable kinetic data from healthy dogs trotting on a treadmill. Ten adult male Malinois Belgian Shepherd dogs were made to trot on an instrumented treadmill to record the ground reaction force for the entire body and to determine the vertical force variables (peak [PFz], impulse [IFz], stride time [Str], peak time [Tz] and contact time [Ct]). Data were collected from each dog, during three sequences per day, on three consecutive days. In order to determine the contribution of the 'sequence', 'day of measurement', and 'dog' factors and the percentage of variance attributable to dogs, data were analyzed with a linear mixed model. The curve shapes were similar to those obtained with a floor-mounted force platform. Intra-clog coefficients of variation were between 1.57 and 3.46%. Inter-dog coefficients of variation were between 4.18 and 7.82%. A sequence effect was not noted. Each day had a significant effect on all of the data. All variables differed significantly from the first day compared to the other days. However there was not any difference between days 2 and 3. The percentage of the total variance attributable to dogs ranged from 37 to 88%. The coefficients of variation were lower than those obtained with common protocols. The treadmill locomotion remained consistent during a single session. Even if interday variation needs to be accounted for, reliable data can still be obtained after a single training session. The majority of the variation was attributable to the dog. An instrumented treadmill may be used for kinetic analysis.