Objectives: we aimed to estimate vaccine effectiveness against severe measles based on the number of vaccine doses administered and the time since last vaccination. Patients and methods: we included measles cases aged at least 2 years and born since 1980 who were notified in France between 2006 and mid-2019. We considered two severity levels (moderate, severe) and calculated adjusted relative risks (aRR) using multinomial logistic regression. Results: we included 10,399 cases. The risk of severe measles in two-dose vaccine recipients was 71% (aRR = 0.29 [95%CI 0.12-0.72]) and 83% (aRR = 0.17 [95%CI 0.04-0.70]) lower than in unvaccinated cases, if the time since last dose was less or more than 15 years, respectively. The risk of moderate disease followed a similar pattern. Conclusions: two-dose measles vaccination provided long-term protection against severe cases, even after vaccine failures. These findings underscore the need for compliance to the recommended measles vaccination schedule to prevent severe cases.