Background: While clinical relapses are the defining feature of relapsing–remitting multiple sclerosis (RRMS), their characteristics vary widely from patient to patient. This study sought to identify predictors of MS relapse location. Based on the current literature, two potential predictors were identified: treatment with interferon β (IFNB) and location of previous relapse. Methods: Patients with RRMS were identified from the UCSF MS Center database who underwent at least 3 months of treatment with IFNB or glatiramer acetate (GA). The relapse immediately preceding the initiation of treatment (pretreatment relapse) and the first relapse occurring after the initiation of treatment (on-treatment relapse) were coded as affecting the spinal cord (SC), optic nerve (ON), brainstem/cerebellum (BC) or cerebrum. Logistic regression was performed to identify independent predictors of on-treatment relapse location. Results: The 134 IFNB and 56 GA patients did not differ in gender, race, age at symptom onset (30.3 years) or disease duration at the start of treatment (5.7 years). Patients with pretreatment SC relapses had increased odds of having on-treatment SC compared with non-SC relapses (OR 2.31, p = 0.013); the same tendency for localisation occurred with BC (OR 3.05, p = 0.013) and ON (OR 3.63, p = 0.011) relapses. Additionally, patients who relapsed on treatment had a higher SC (but not ON or BC) relapse risk when they were receiving IFNB compared with GA (OR 2.05, p = 0.041), independent of pretreatment relapse location. Conclusion: These results show a tendency for patients to have localised exacerbations, which could be mediated by genetic or immunological factors. In addition, and to be confirmed in subsequent studies, IFNB treatment may influence SC relapse risk.