This paper addresses the implementation of a predictive control strategy for Un-manned Air Vehicles in the presence of bounded disturbances. The goal is to prove the feasibility of such a real-time optimization-based control design and to demonstrate its tracking capabilities for the nonlinear dynamics with respect to a reference trajectory which is pre-specified via differential flatness. In order to benefit from the computational advantages of the linear predictive control formulations, an off-line linearization strategy of the nonlinear model of the vehicle along the flat trajectory is employed. The proposed method exhibits effective performance validated through software-in-the-loop simulations and real flight tests on different Unmanned Aerial Vehicles (UAVs).