To investigate: (a) the interrater and test-retest reliability of the 3-m backward walk test (3MBW) in ambulant people with multiple sclerosis (PwMS); (b) minimal detectable change (MDC); (c) concurrent and known-groups validity; and (d) the cutoff time to best discriminate fallers from nonfallers with multiple sclerosis (MS). Forty-nine PwMS and 36 healthy people were included in this cross-sectional study. The 3MBW was administered with the timed up and go test, Berg Balance Scale, four square step test, Falls Efficacy Scale-International, and Expanded Disability Status Scale. The 3MBW was simultaneously performed by two independent raters to examine the interrater reliability while was repeated after 7-10 days to examine the test-retest reliability. The 3MBW showed good interrater reliability [intraclass correlation coefficient (ICC) = 0.987-0.989] and excellent test-retest reliability (ICC = 0.854-0.889). The MDC was found to be 1.69 s. The 3MBW had moderate-to-strong correlations with the other measures. For the 3MBW, PwMS had worse performance than healthy people (P < 0.001), whereas fallers with MS had worse performance than nonfallers with MS (P < 0.001). The 3MBW time of 7.86 s was determined to best discriminate fallers from nonfallers with MS. The 3MBW is a reliable, simple, and easy-to-administer tool for assessing backward walking among ambulant PwMS.