This paper presents a full dynamic visual servoing flight controller design using command filtered backstepping (CFBS) control law for a fixed-wing unmanned aerial vehicle (UAV). A full nonlinear dynamic model that involves feature point motion, pan-tilt gimbal mechanism movement, and UAV flight dynamic is adapted to the controller design of CFBS. The proposed design scheme can provide a system which is easy to implement in various fixed-wing UAV platforms, since it only needs physical characteristics such as mass, mass of inertia, center of gravity, geometry and propeller-engine to configure the system. Further, additional novel algorithms are developed and added to the original CFBS control law design to make longitudinal and lateral-directional maneuvers safer and smoother. The proposed algorithm is implemented and tested in both numerical simulation and hardware-in-the-loop simulation (HILS). HILS is conducted by implementing the algorithm on the real UAV on-board hardware that consist of an embedded PC for image extraction and a microcontroller for the flight controller. The numerical simulation and HILS results show that the developed system is able to perform target tracking and pursuing task effectively.