In order to improve the operational effectiveness and operational indicators of the UAV (unmanned aerial vehicle), and to enhance the accuracy and reliability of the UAV relative navigation system, a novel relative navigation method was proposed. Under the background of relative navigation system, the cubature information filter based on the cubature Kalman filter and information filter was studied. Moreover, an INS/GPS/VisNav relative navigation filter was designed by making use of the multi-sensor information fusion theory and distributed information fusion structure to fuse the information from INS, VisNav and GPS, and then the relative position, velocity and attitude were obtained. By making use of this algorithm, the accuracy, reliability and stability of the reliability navigation system were all improved. In addition, the accuracy decrease and numerical instability which often occur to traditional filter were avoided by cubature information filter. Mathematical simulation results indicate that the method can improve the accuracy and reliability of the UAV relative navigation system, and the proposed algorithm is verified.