Abstract: When strong jamming or signal is presented,its peak in the spatial spectral will submerge the peak of signals with lower power.For signal direction-of-arrival(DOA) estimation in the presence of strong jamming or signals,an idea of modifying the noise subspace and then a new method is presented.First,to form the extended noise subspace,the traditional noise subspace is modified by adding the space composed of the steering vectors of strong signals into it,then conventional(multiple signal classification,MUSIC) is applied to this extended noise subspace and the DOAs of low SNR(signal-to-noise ratio) signals are obtained.The peaks of strong signals in spatial spectrum are attenuated significantly and effectively by extending the noise subspace and it does not require either the angle of incidence of strong signals or the steering vectors of them as preknowledge and has less computation burden compared with the other algorithms.Theoretic analysis shows the DOA estimation performance of this new algorithm is not less than its corresponding blocking algorithms,and computer simulation demonstrates its effectiveness and feasibility.