Leading PV system operation at its Maximum Power Point (MPP) is crucial to improve the overall system efficiency and reduce the solar energy cost. In such a sense, this paper presents a closed-loop Maximum Power Point Tracking (MPPT) method based on the combination of a Second Order Sliding Mode Observer (SOSMO) and a traditional PI controller. The particular SOSMO proposed is especially appropriate to deal with finite time power gradient estimation even in the presence of unmodeled dynamics and external perturbations. Based on this estimation, the PI controller is used to quickly drive the PV system to operate on its MPP and tracking its variations. The resulting closed -loop scheme presents a fast tracking dynamics during rapidly changing environmental conditions having good accuracy and a moderate computational burden. The closed -loop system performance is evaluated using representative computer simulations and corroborated through experimental tests.