A novel method for high resolution direction-of-arrival (DOA) estimation is proposed. The novel approach constructs the spatial spectrum by utilizing the steady-state weights of the adaptive algorithm in direct data domain instead of the noise subspace. The proposed method can solve the false peak problem effectively in the spectral peak search. The noise subspace can be closed with several groups of linear non-correlated steady-state weights from utilizing different reference signals. A practical method for the DOA estimation of correlated signals is also presented, which utilizes the symmetric conjugate vector in direct data domain. Compared with the conventional method, the proposed approach does not need to estimate the source number and the covariance matrix. As it does not need the eigendecomposition and only has a computational complexity of O(MP), it is a practical method for DOA estimation of correlated signals.