An efficient STAP algorithm for NonSideLooking (NSL) airborne radar is presented. The algorithm can mitigate the range dependence of clutter by mainlobe clutter compensation. To reduce the computational complexity, the Doppler frequency of the mainlobe clutter is firstly estimated via FFT in the time domain, and then the spatial frequency is accurately obtained by sparse reconstruction corresponding to the output of the mainlobe clutter Doppler cell. Therefore, based on the estimated location of the mainlobe clutter, the clutter corresponding to different range cells can be adaptively compensated, which results in improved clutter suppression performance of the following 3DT processing. As shown in the simulation, the Improvement Factor (IF) of 3DT is increased about 18 dB in the domain of mainlobe, which can greatly improve the detection of slowmoving targets. In addition, the proposed scheme can be applied to real-time processing owing to its small computational load.