Abstract: Airborne radar with uniform circular array antennas has the merits of flexible search and track mode, all-orientation space scanning and good beamform when the beam center deviates from the normal of the array.In this paper, the signal model of uniform circular array is set up and the clutter characteristics of uniform circular array are analyzed from the aspects of distribution of clutter spectral and the distribution of eigenvalues of the clutter matrix. A new space time adaptive processing method for uniform circular array, namely localized cross spectral metric algorithm, is proposed. In this method, structured reduced dimension technology is firstly adopted to reduce the dimension of processer and then cross spectral method is used to select N eigenvectors of clutter matrix which correspond with the N maximum cross spectral metric to construct the secondly reduced dimension matrix. Simulation results show the proposed method outperforms significantly localized processing methods while the computation load and independently identically distributed samples needed are about the same as those methods and is robust when the error occurs.