An adaptive control scheme for maximum power point tracking of a single-phase grid-connected photovoltaic system is presented. The dificulty on design a controller that may operate a photovoltaic system on its maximum power point (MPP) is that, this MPP depends on temperature and solar irradiance, ambient conditions that are time-varying and dificult to measure. A solution using an on-line sliding mode estimator is presented. It estimates three different parameters that depends on solar irradiance and temperature, eliminating the necessity of having any sensor for these environmental variables. It is capable of estimate time-varying parameters. A complete analysis was done taking into account the non-linearities showed by the closed-loop system. A variation of the Lyapunov redesign technique was used to find a controller that give globally asymptotically stable trajectories of the closed-loop system. An adaptive law was found to substitute a perturbation bound and also to eliminate possible chattering due to the discontinuous controller term. Computer simulations are presented to show the good performance of the controller.