Long-term demographic surveys, needed to obtain accurate information on population dynamics and efficiently manage rare species, are still very scarce. Matrix population models are useful tools to identify key demographic transitions and thus help setting up conservation actions. Furthermore, the combination of ecological, demographic and genetic data is likely to improve the identification of the threats acting upon populations and help conservation decisions. In this paper we illustrate the power of this approach on Brassica insularis, a Mediterranean endemic plant species, rare and endangered in Corsica (France). In four populations of this species, a long-term demographic survey (2000)(2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009), genetic analyses (in 2000 and 2009) and survey of ecological variables (climatic variables, competition and herbivory) were performed. By using both deterministic and stochastic matrix model analyses, we assessed the viability of each population and tested for both spatial and temporal variations in demographic vital rates. Populations exhibited differing demographic behaviours and environmental stochasticity occurred in populations. Significant correlations between climatic variables and vital rates were detected. Stochastic simulations suggested that three out of the four populations studied might present a high risk of extinction on the short-term and should actively be managed, or at least surveyed. It could be, however, that two of these populations are experiencing density-dependent regulation, rather than being declining. Microsatellite diversity was slightly reduced in a single population and similar in the three others, consistently with expectations based on population census size and geographic area, as well as with diversity at the S-locus observed in 2000. The combination of all data led to specific recommendations for managing each population. We discuss the implications for conservation of such a general approach.