This paper proposes direct blind equalization corresponding to noisy environment using Rayleigh Quotient. Direct blind equalization using Zero-Forcing (ZF) criterion as the conventional method has problem that the performance of equalizer is lower because this method is used for the signals including noises. To solving this problem, in this paper, we propose Direct Blind Equalization based on Data least Square (DLS) method consideration on the noises. The proposed method is directly estimated equalizer by observed signals for modeling input-output relation of equalizer based on constrain conditions of ZF criterion and the matrix constructed observed signals is input signals of systems. The noises of the matrix constructed observed signals is suppressed DLS method based on Rayleigh Quotient. The features of proposed method are (i) realization of high performance and stable equalizer estimation with Rayleigh Quotient and (ii) construction of new model for suppressing the noises. We show the effectiveness of the proposed method using computer simulation.