An autonomous navigation scheme for unmanned aerial vehicles (UAVs) is presented innovatively. Navigation information can be given by integrated navigation system without the known land-based synergic targets. An optical flow method is adopted by speed-up robust features (SURF) algorithm in the visual navigation. A multi-rate indirect and feedback-error-correction-based Kalman filter is built to integrate the measurement data of strapdown inertial navigation system (SINS), electronic compass, optical flow, and altimeter system and laser range finder. And the estimations of attitude angles, altitude, and horizontal velocity of UAV can be given. When it is not preferable to build the manmade land-based synergic targets for computer vision equipment of UAV, the navigation scheme can be used such as the low altitude flying phase and the approach phase. The autonomous navigation scheme can run accurately in the simulation.