Abstract: The performance of traditional anti-jamming methods deteriorated under the condition of low signal-to-noise ratio (SNR). In order to improve the anti-jamming ability of radar under the condition of low SNR, in this paper, we proposed an anti-jamming method combining subarray processing and sparse recovery. In this paper, subarray Adaptive Digitial Beam Forming (ADBF) and sparse recovery are used to suppress the mainlobe interference and sidelobe interference. Firstly, the whole array was divided into several subarrays according to certain rules. Then, the ADBF technique was used to suppress the sidelobe interference signals of each subarray, and the output data of each subarray were combined to construct new array data. Finally, the sparse representation was constructed from the newly constructed array data, the Reduce Multiple measurement vectors and Boost (ReMBo) was improved to reduce computation in the condition of low SNR, and the Alternating Direction Method of Multiplier (ADMM) was used to perform the sparse recovery, the target signals were separated and the Direction of Arrival (DOA) estimation was obtained. Simulation results show that the proposed method can effectively suppress the mainlobe and sidelobe interference, and this method has better performance in target detection and DOA estimation compared with other methods, especially under the condition of low SNR. Under the condition of low SNR, the new method can effectively suppress the mainlobe and sidelobe interference and achieve better results than traditional methods.