Objective: To investigate the bivariate correlations between objective physical activity (PA), self-efficacy, and a range of multiple sclerosis (MS) symptom measures. Also, to determine whether any MS symptom measures moderate the relationship between self-efficacy and PA. Method: Baseline analysis from a randomized control trial exercise plus a behavior change intervention was used. Fifty-nine physically inactive people with a definite diagnosis of MS were included. Participants were asked to record 7-day objective PA with the SenseWear Armband. Additionally, measures of self-efficacy (Exercise Self-Efficacy Scale), walking endurance (6-min walk test), impact of walking (12-Item MS Walking Scale [MSWS-12]), anxiety and depression (Hospital Anxiety and Depression Scale [HADS]) and fatigue (Modified Fatigue Impact Scale) were assessed before intervention. Pearson’s correlations and moderation analysis were conducted. Results: Self-efficacy was positively associated with PA (r = .30, p < .05). HADS-Depression (r = −.25, p < .05) and MSWS-12 (r = −.31, p < .01) demonstrated statistically significant negative correlations with self-efficacy but not with PA. Moderation analysis illustrated a significant interaction between anxiety and self-efficacy (r = .39, p < .03) with an R2 value of .15. The interaction was significant at lower values of anxiety, suggesting that the relationship between self-efficacy and PA is stronger when levels of anxiety are lower. Conclusion: This article confirms the potential role of MS symptoms, in particular anxiety, in explaining PA behavior in an inactive sample of persons with multiple sclerosis (pwMS). Further exploration is warranted, and future PA interventions should acknowledge the potential interplay of psychosocial constructs such as self-efficacy and anxiety in changing PA behavior among pwMS.