In view of recognition of interrupted-sampling repeater jamming, this paper puts forward a method for jamming recognition based on sparsity in frequency domain, by taking advantage of the difference between the signal spectra of jamming and target echoes. By implementing the Fourier Transform to three types of interrupted-sampling repeater jamming and the target echo signals, it mines the difference information of sparsity in signal spectrum, extracts the spectrum sparseness to be the characteristic parameters of jamming recognition,and makes use of the SVM to perform the classification and recognition. Simulation results show that this proposed method can implement the signal recognition of different types of Jamming and target echoes, which has a higher recognition rate, thus offering a reference for the radar system taking a targeted anti-jamming measures.