This paper discusses energy harvesting by small Uninhabited Aerial Vehicles (UAV s). A receding horizon controller which computes a sequence of pitch rate commands with the goal of maximizing energy gain over a fixed horizon is derived. An energy based reward function is used to maximize energy gain with only local knowledge of atmospheric wind conditions. Terms are included in the developed reward function to drive the aircraft towards steady-state ight at the end of every plan horizon. The coefficients in the reward function are tuned by the used of an evolutionary algorithm. The controller developed is used in simulated ight through steady winds, Dryden gust fields at different simulated altitudes and intensities, and through random thermal fields. The results show that the controller is effective in maximizing energy gained from the surrounding air, resulting in altitude or velocity gain. The majority of results compare favorably to a constant speed controller. A study of the computation time for this method is also presented to assess the practicality of application.