Compared with the conventional pulse Doppler radar signal processingSpace-Time Adaptive Processing (STAP) expands the signal processing dimensionso that the clutter and the target can be distinguished in the joint space-time domain.Based on sparse representation theory and the sparsity of clutter spectrumSparse Recovery (SR)-based STAP realizes clutter suppression under the condition of a small number of training range cells.Aiming at the performance degradation of SR-STAP methods with unknown yaw anglea nonlinear regression-based clutter reconstruction STAP method is proposed.Firstlybased on SR clutter spectrumoutlier degree is used as the convergence objective to iteratively eliminate the scatter points deviating from the ridgeand the coordinate weighted nonlinear regression is performed to realize accurate estimation of the parameters of the clutter ridge model.Thenbased on the results of the first screeningthe clutter spectrum is estimated accurately by the nonlinear regression method again.Finallythe clutter reconstruction and suppression are completed based on the above estimation results.The simulation verifies the effectiveness of the proposed methodand compared with the existing STAP methodsit achieves better space-time frequency response and SINR losseffectively improving the performance of clutter suppression and moving target detection.