The aim of the study was to establish the test-retest reliability, clinical significance and precision of four mobility and balance measures--the Timed 25-Foot Walk, Six-minute Walk, Timed Up and Go and the Berg Balance Scale--in individuals moderately affected by multiple sclerosis. Twenty four participants with multiple sclerosis (Extended Disability Status Score 5-6.5) were assessed on four measures of mobility and balance. The Timed 25-Foot Walk, Six-minute Walk and Timed Up and Go mobility outcome measures and the Berg Balance Scale were assessed by one assessor one week apart. Intraclass correlation coefficient (ICC) analysis was carried out to determine reliability. Minimal detectable change values were calculated to determine clinical significance; the standard error of each measurement was calculated to assess precision. All four outcome measures were found to be reliable: Timed 25-Foot Walk ICC = 0.94, Six-minute Walk Test ICC = 0.96, Timed Up and Go ICC = 0.97 and Berg Balance Scale ICC = 0.96. Minimal detectable change values were as follows: Timed 25-Foot Walk = 12.6 s, Six-minute Walk Test = 76.2 m, Timed Up and Go = 10.6 s and Berg Balance Scale = 7 points. Standard errors of measurement were as follows: Timed 25-Foot Walk = 4.56 s, Six-minute Walk Test = 27.48 m, Timed Up and Go = 3.81 s and Berg Balance Scale = 3 points. The test-retest reliability of these four outcome measures was found to be good. The calculated clinical significance and precision of these measures highlight the problems of assessing a heterogeneous clinical population.