In modern warfare, the battlefield environment is very complicated. Automatic and intelligent jamming is the main threat for the precise detection of radar systems. Methods based on freedom in array spatial domain can suppress the side-lobe jamming effectively. However, for the main-lobe jamming, traditional anti-jamming methods fail to reduce the negative impact. Under the interfering circumstances, the target’s Signal to Noise Ratio(SNR) is low. Lower detection threshold will cause more false alarm points. In this paper, a detection technique based on sparse representation is proposed for small target detection under jamming of main lobe noise. This detection method uses the characteristics that the target can be represented sparsely but the noise cannot, so as to suppress the noise and reduce the false alarm rate. The simulation results show the effectiveness of the proposed method.