Abstract: Aiming at the problem that the traditional radar interception beam pointing method is only suitable for uniform speed operation, in order to solve the interception requirements for the operation during variable speed and non-linear marching, a fusion positioning method of lidar and inertial navigation using kernel adaptive filtering algorithm is proposed. Compared with the traditional method of inertial navigation speed compensation, the proposed method has such characteristics as high positioning accuracy and good robustness, and it can improve the beam pointing accuracy, shorten the interception time and improve the combat efficiency of the weapon system, and it has certain research value for missile interception during radar marching in complex scenes.