An approach to estimation for a class of hybrid discrete-time linear systems using Luenberger observers is presented. The proposed Luenberger observer for such a kind of systems relies on the switching among different gains. Convergence conditions have been found to ensure the stability of the error dynamics and the related gains may be selected by solving a set of linear matrix inequalities (LMIs). Moreover, this observer may be improved by suitably updating the estimate using the last measures. This update enables one to reduce the norm of the estimation error and is based on the so-called projection method. Simulation results are reported to show the effectiveness of these methods in the estimation for hybrid discrete-time linear systems.