Introduction: There has been increased interest in the objective monitoring of free-living walking behavior using accelerometers in clinical research involving persons with multiple sclerosis (MS). The current investigation examined and compared the accuracy of the StepWatch activity monitor and ActiGraph model GT3X+ accelerometer for capturing steps taken during various speeds of prolonged, over-ground ambulation in persons with MS who had mild, moderate, and severe disability. Methods: Sixty-three persons with MS underwent a neurological examination for generation of an EDSS score and undertook two trials of walking on the GAITRite electronic walkway. Participants were fitted with accelerometers, and undertook three modified six-minute walk (6MW) tests that were interspersed with 10-15 minutes of rest. The first 6MW was undertaken at a comfortable walking speed (CWS), and the two remaining 6MW tests were undertaken above (faster walking speed; FWS) or below (slower walking speed; SWS) the participant's CWS. The actual number of steps taken was counted through direct observation using hand-tally counters. Results: The StepWatch activity monitor (99.8%-99.9%) and ActiGraph model GT3X+ accelerometer (95.6%-97.4%) both demonstrated highly accurate measurement of steps taken under CWS and FWS conditions. The StepWatch had better accuracy (99.0%) than the ActiGraph (95.5%) in the overall sample under the SWS condition, and this was particularly apparent in those with severe disability (StepWatch: 95.7%; ActiGraph: 87.3%). The inaccuracy in measurement for the ActiGraph was associated with alterations of gait (e.g., slower gait velocity, shorter step length, wider base of support). Conclusions: This research will help inform the choice of accelerometer to be adopted in clinical trials of MS wherein the monitoring of free-living walking behavior is of particular value. © 2014 Sandroff et al.