In order to achieve autonomous behavior of UAVs in surveillance missions, it is of interest to embed a vision sensor as an integral part of the control loop. The major difficulty in this type of applications is the lack of realistic simulation environments. Those, if available, can avoid the obvious dangers and cost of extensive field experimentation with visual servoing controllers for airborne platforms. This paper investigates the performance and possibilities of various visual-servoing techniques applied to acquisition and tracking in the case of UAVs. The work is done in SEAVS, a 3D simulation environment for aerial visual-servoing. SEAVS is a simulation software with a graphical user interface including one 2D window for planning of the aerial mission and two 3D windows -- one for the visualization of the ongoing simulation, and one showing the present camera view. The unique feature of SEAVS is the use of orthographic photos which enhances realistic image processing. The paper describes the design of visual-servoing controllers for the purpose of acquisition and tracking tasks. Two types of controllers are investigated: one based directly on the angle errors and one based on the image Jacobian. The algorithms are validated by simulation in SEAVS.