As part of the synthesis of a tracking system for an unmanned aerial vehicle (UAV) exposed to uncontrolled external disturbances under the conditions of incomplete measurements of the state vector, we develop procedures for synthesizing novel low-order state and disturbance observers that do not require constructing dynamic models of external inputs. The observation subsystem includes two state observers. One of them estimates the velocities based on measurements of the UAV center of mass coordinates. The other observer uses the measurements of tracking errors to give estimates of mixed variables (state functions, external inputs, and their derivatives) from which the feedback is directly formed. It is noted that implementing the developed algorithms, which do not involve readjustment when external inputs change, will increase the UAV control system functionality and its reliability in the event of failure of measuring devices. The efficiency of our approach to tracking system synthesis is confirmed by numerical simulation results. We present the results of comparative analysis of closed-loop systems with static (under the assumption that all internal and external variables are measured) and dynamic feedback that uses two approaches to solving the problem of estimation under external disturbances—high-gain observers and observers with piecewise linear bounded corrective inputs. It is shown that, despite the simpler setup, it is expedient to use observers of the second type in linear feedback systems, while high-gain observers will be in demand in systems with a priori bounded controls.