Because of the deficiency of the same type of radar in extracting micro-motion parameters of ballistic target, a method for precession feature extraction of ballistic target was proposed by using hybrid-scheme radar network under the circumstance that the numbers of the obtained scattering centers are different. First, a ballistic target precession model and a wide band and narrow band radar signal model were established, and the echoes characteristics of the target scattering centers under narrow band radar and wide band radar were analyzed. Based on the characteristics, each scattering center in different radar system echoes was matched and identified. Then the scattering center range information obtained by the wide band radar was analyzed, the range information was transformed by using generalized Radon transform, and the precession angle was solved by parameter transform relationship. The associated systems of equations of micro-Doppler information were further established, and the visual angle and structural parameters of radars were solved by the precession angle. Dichotomy was used to optimize the parameters obtained from the three solution combinations, and the obtained parameters were constructed to obtain three-dimensional cone rotation vector. Simulation results show that when the SNR was greater than 5 dB, the validity of the proposed method was close to that of the wide band network and both were higher than that of the narrow band network, which verified the effectiveness of the proposed method.