Abstract: The estimation accuracy of direction of arrival (DOA) in the incoherent signal subspace method (ISSM) under low signal-to-noise ratio condition is susceptible to the interference from data containing estimation errors of DOA angles at different frequency points. Furthermore, the ISSM method cannot solve the DOA estimation problem of broadband coherent signals. To overcome the limitations of the method, a novel broadband coherent DOA estimation method by building upon and refining the existing ISSM approach is proposed in this paper. Firstly, the frequency cross-spectral density matrix of different frequency points is obtained by the ISSM method, and the de-coherence processing is carried out. Then, the subspace DOA algorithm is used to calculate the DOA estimation results of each frequency point. Finally, the interval density method is used to fuse the DOA estimation results of all frequency points to obtain the final DOA estimation result. The simulation results show that the proposed method, which is easy to implement in engineering, not only can effectively solve the estimation problem of wideband coherent signal, but also has better estimation performance and higher estimation accuracy under low signal-to-noise ratio.