This paper presents a flight control system designed as on-board intelligence for rotorcraft-based unmanned aerial vehicles (RUAVs). This hierarchical flight control system, endowed with autonomy resembling sense-reason-act processes of intelligent agents in nature, gradually refines given abstract mission commands into real-time control signals for each vehicle. A tracking control layer is designed on the identified vehicle dynamics and integrated with a trajectory generator for logistical action planning. The proposed structure has been implemented on radio-controlled helicopters and validated in a variety of experiments. Results from way-point navigation, a probabilistic pursuit-evasion game and vision-based tracking of a moving target show the promising potential of intelligent flying robots.