This paper investigates the problem of uncooperative target tracking for unmanned aerial vehicles (UAVs) using vision-based measurements. We propose a novel control strategy that enables a follower UAV to accurately track a target while ensuring that both the control inputs and resulting velocities remain within prescribed bounds. The proposed algorithm is structurally simple yet robust, effectively handling model uncertainties, external disturbances, and physical constraints. By appropriately selecting design parameters and leveraging Lyapunov-based analysis, we demonstrate that asymptotic tracking is guaranteed. Extensive simulation validation verifies the efficacy of our theoretical results under various operating conditions, which provides practical insights for UAV control applications.