MARS (Maintenance Assessment of Repairable Systems) takes as input data sets consisting of successive times of preventive and corrective maintenance performed on repairable systems, possibly left and right-censored. It implements the main imperfect maintenance models, of the virtual age type. It estimates the parameters of these models by the maximum likelihood method. This allows to assess the main reliability indicators such as ageing speed, maintenance efficiency, mean time to failure or mean cumulative number of failures on a given period of time. It plots and saves different graphs such as the reliability function, failure intensity or the men cumulative number of failures. It has features of extensive simulation in order to study the statistical properties of estimates and to make an accurate prognosis on the maintenance process.