Abstract: By introducing cyclostationary theory into the estimation of direction of signal of arrival, a novel extended cyclic MUSIC algorithm based on uniform circular array was developed. The proposed method could eliminate both noises and interferences in the same frequency band by exploiting simultaneously the information contained in both cyclic correlation matrix and cyclic conjugate correlation matrix of the extended array data vector. It has the ability to select desired signals and ignore interferences from the received data. Simulation results demonstrate that the proposed algorithm has a high DOA estimation accuracy and multi-signal resolution. On the other hand, the proposed method also breaks through the limitation of the classic MUSIC algorithm in which the number of source signal cannot be greater than the number of element in the receiving array.