Recent developments in autopilot technology have increased the potential use of micro air vehicles (MAVs). As the number of applications increase, the demand on MAVs to operate autonomously in any scenario increases. Currently, MAVs cannot reliably fly in cluttered environments because of the difficulty to detect and avoid obstacles. The main contribution of this research is to offer obstacle detection and avoidance strategies using laser rangers and cameras coupled with computer vision processing. In addition, we explore methods of visual target tracking and task allocation. Utilizing a laser ranger, we develop a dynamic geometric guidance strategy to generate paths around detected obstacles. The strategy overrides a waypoint planner in the presence of pop-up-obstacles. We develop a second guidance strategy that oscillates the MAV around the waypoint path and guarantees obstacle detection and avoidance. Both rely on a laser ranger for obstacles detection and are demonstrated in simulation and in flight tests. Utilizing EO/IR cameras, we develop two guidance strategies based on movement of obstacles in the camera field-of-view to maneuver the MAV around pop-up obstacles. Vision processing available on a ground station provides range and bearing to nearby obstacles. The first guidance law is derived for single obstacle avoidance and pushes the obstacle to the edge of the camera field-of-view causing the vehicle to avoid a collision. The second guidance law is derived for two obstacles and balances the obstacles on opposite edges of the camera field-of-view, maneuvering between the obstacles. The guidance strategies are demonstrated in simulation and flight tests. This research also addresses the problem of tracking a ground based target with a fixed camera pointing out the wing of a MAV that is subjected to constant wind. Rather than planning explicit trajectories for the vehicle, a visual feedback guidance strategy is developed that maintains the target in the field-of-view of the camera. We show that under ideal conditions, the resulting flight paths are optimal elliptical trajectories if the target is forced to the center of the image plane. Using simulation and flight tests, the resulting algorithm is shown to be robust with respect to gusts and vehicle modeling errors. Lastly, we develop a method of a priori collision avoidance in assigning multiple tasks to cooperative unmanned air vehicles (UAV). The problem is posed as a combinatorial optimization problem. A branch and bound tree search algorithm is implemented to find a feasible solution using a cost function integrating distance traveled and proximity to other UAVs. The results demonstrate that the resulting path is near optimal with respect to distance traveled and includes a significant increase in expected proximity distance to other UAVs. The algorithm runs in less than a tenth of a second allowing on-the-fly replanning.