The paper deals with tracking systems for unmanned aerial vehicle under external disturbances. Within the framework of the block approach, the combined control law is developed that provides invariance with respect to external disturbances in a closed system via linear control with saturation. Under the assumption that the analytical form of the target signals is not known, dynamic differentiators with piecewise linear correcting influences are designed to estimate unknown derivatives of target signals. These differentiators are replicas of canonical virtual models with unknown inputs and provide estimates of the derivatives of deterministic signals with a given accuracy in a given time. The simulation results are presented, confirming the effectiveness of the developed algorithms.