Abstract Background Rare diseases (RDs) encompass a heterogeneous group of life-threatening or chronically debilitating conditions that individually affect a small number of subjects but overall represent a major public health issue globally. There are still limited data on RD burden due to the paucity of large population-based epidemiological studies. The aim of this research was to provide survival estimates of patients with a RD residing in Tuscany, Italy. Methods Cases collected in the Rare Diseases Registry of Tuscany with diagnosis between 1st January 2000 and 31th December 2018 were linked to the regional health databases in order to retrieve information on mortality of all subjects. Survival at 1, 5 and 10 years from diagnosis with 95% confidence intervals (CI) was estimated by sex, age class, nosological group and subgroup using the Kaplan–Meier method. The effect of sex, age and period of diagnosis (years 2000–2009 or 2010–2018) on survival was estimated using Cox proportional hazards regression. Results Survival at 1, 5 and 10 years from diagnosis was 97.3%, 88.8% and 80.8%, respectively. Respiratory diseases and peripheral and central nervous system disorders were characterized by the lowest survival at 5 and 10 years. Despite a modest higher prevalence of RDs among females (54.0% of the total), male cases had a significant increased risk of death (hazard ratio, HR 1.48, 95% CI 1.38–1.58). Cases diagnosed during 2010–2018 period had a risk of death significantly lower than those diagnosed during 2000–2009 (HR 0.81, 95% CI 0.82–0.96), especially for immune system disorders (HR 0.48, 95% CI 0.26–0.87), circulatory system diseases (HR 0.61, 95% CI 0.45–0.84) and diseases of the musculoskeletal system and connective tissue (HR 0.64, 95% CI 0.49–0.84). Conclusions An earlier diagnosis as well as the improvement in the efficacy of treatment resulted in a decreased risk of death over the years for specific RDs. The linkage between a population-based registry and other regional databases exploited in this study provides a large and accurate mass of data capable of estimating patients’ life-expectancy and increasing knowledge on the collective burden of RDs.