The problem of autonomous landing of unmanned aerial vehicle（ UAV） under the aid of visual navigation is studied. Use image recognition method such as Harris corner detection and contour detection

pose estimation algorithm based on pinhole camera model is designed. Autonomous landing control algorithm for UAV is studied. The pose controller and the pan/tilt/zoom（ PTZ） controller in the process of UAV landing are designed using proportional-integral-differential（ PID） control method. On this basis

three landing modes of UAVs are innovatively proposed

and the autonomous landing test is carried out in each mode. The test results show that the UAV can be successful landing in three modes. Due to low visual feedback frequency

UAV in the first two modes has a larger landing error. UAV in the third mode achieves a high-precision landing because of using PTZ servo tracking control and speed control.