Abstract: Aiming at the problem that the fourth-order cumulant has high computational complexity in multiple signal classification (MUSIC) and poor performance under low signal-to-noise ratio conditions, a modified fourth-order cumulant propagator method (MFOC-PM) was proposed. By introducing the propagation operator algorithm to obtain the noise subspace linearly, the algorithm avoided the computation caused by eigenvalue decomposition. To maintain the advantage of the aperture expansion of the fourth-order cumulant array and suppress Gaussian noise, the redundant part of the fourth-order cumulant was removed by the method of averaging and then removing. The experimental results demonstrate that the effectiveness and feasibility of the proposed MFOC-PM algorithm, which significantly reduces the computational complexity of the algorithm. Specifically, it achieves accurate angle estimation and performs well under low signal-to-noise ratio conditions. This indicates that the algorithm has significant practical value and can meet the requirements of DOA estimation in many practical application scenarios.