The operation of manned and unmanned aerial vehicles, over populated areas and inside non-segregated airspaces, imposes the need of coordinated flight among vehicles and with the air traffic control. In case of an eventual loss of communication with the groundbased traffic control, or lack of visual contact with another aircraft, the operator must autonomously provide safe separation and avoid the collision. While this procedure is normally employed by manned aircraft pilots, unmanned aerial vehicles currently are not able to see or detect and avoid another conflict traffic. Thus, this work presents a new method to provide three-dimensional guidance in order to perform, autonomously, collision avoidance and also to restore adequate safe separation, without any type of on-line optimization algorithms. The proposed method is analyzed through several possible encounter scenarios, showing adequate performance with little computational burden. Nomenclature χ Course, rad ∆χ Course error γ Flight path angle, rad κ Path curvature λ Longitude, rad µ Latitude, rad Ω Turn rate, rad/s ω n Natural frequency ζ Damping ratio D Horizontal distance, m D N Distance in the north direction, m D E Distance in the east direction, m D D Distance in the down direction, m D min Minimum horizontal distance D saf e Safe distance e L Lateral distance error H Vertical distance, m H min Minimum vertical distance h Altitude, m K LAT Lateral guidance gain matrix K LON Longitudinal guidance gain matrix R E Earth radius (6378137 m) R t Path turn radius T min Minimum safe TTE value V True airspeed, m/s V G Ground (horizontal) speed, m/s V Z