OBJECTIVES:  To study the long-term overall survival (OS) with MS and its modifiers in Finland. METHODS:  Proportionate risks of death were assessed from MS onset to death or up to 31.12.2010 in a cohort consisting of incident relapsing onset (RRMS) and primary progressive (PPMS) groups, and in subgroups of DMT exposed and non-exposed RRMS patients. MS diagnosis was determined by definite Poser criteria from 1.1.1981 to 31.12.2010. The overall survival (OS) and life expectancy by gender and disease course were assessed with Kaplan-Meyer method and log rank test. Hazard ratios (HR, 95%CI) for prognostic factors and confounders were estimated using Cox’s exact marginal likelihood multivariable proportional hazards model. Somers' D and Harrell's C statistics and their 95%CIs were calculated to assess the predictive accuracy of Cox model. RESULTS:  The 25-year OS up to 31.12.2010 was 84% (N=1617, 164 deaths); 93% in DMT treated (n=870); 81% in non-exposed (n=561); and 76% in PPMS (n=189) (log-rank p<0.001) group. Median OS from birth was 78 years. Among the DMT exposed, the proportionate survival benefit was multivariate independent and over time over two-fold (HR 2.3, 1.4–3.7) as compared to non-exposed RRMS group, and concerned both genders across the wide diagnostic timeframe. The OS trend associated with DMT exposure was relatively stable over time and proportional hazards assumption was not violated. Significant adverse OS was related to male gender, PPMS course, increasing age at onset, diagnostic delay and multiple onset symptoms. Cox model's predictive accuracy was good (C 0.76, 0.72–0.80; D 0.58, 0.47–0.69) and the model gained significant log likelihood. CONCLUSIONS:  MS prognosis is highly individual, but significant and stable OS advantage relates to RRMS and DMT use while significant disadvantage relates to male sex, PPMS, diagnostic delay, higher age and multiple symptoms at onset. Life expectancies in the incident MS cohort and Finnish general population were comparable.