In the field of spaceborne long-distance non-cooperative target ranging, usually the target is weak and the imaging frame rate is limited, and thus the target tracking ability is limited and influences the ranging distance. When the spaceborne long-distance non-cooperative target is tracked, the main source of tracking errors is body vibration. In order to enhance the active vibration suppression ability and improve the tracking accuracy of the system, adaptive control is used to realize aiming control of targets. Aiming at the problem of difficult data fusion among different frequency sensors, we use down-sampling to achieve adaptive fusion of low-frequency camera data and high-frequency inertial sensor data, and simultaneously confirm the effectiveness and correctness of adaptive filtering under down-sampling. At the level of algorithm implementation, an affine projection algorithm is selected to achieve data fusion considering the problems of slow convergence and poor steady-state accuracy of the classical adaptive algorithm. In the data fusion simulation, the micro-vibration model of the satellite is used to verify the convergence speed and the tracking accuracy of the fusion method. Finally, this method is applied for the actual tracking system and the experimental results show that the tracking accuracy is superior to 5 μrad, which meets the tracking accuracy requirement of long-distance non-cooperative target ranging.