Abstract: In order to solve the problem of low recognition of signal modulation in low SNR condition,TFC-KNN algorithm was proposed.It was based on Higher-Order Cumulants and K-NN Nearest Neighbor algorithm. According to the algorithm,the Higher-Order Cumulants is introduced on the basis of time frequency analysis,and the signal will be classified using K-NN Nearest Neighbor algorithm.The feature parameter,presented in the algorithm,can effectively restrain the additive white Gaussian noise.Simulation results show that the recognition rate for this recognition structure and algorithm can reach 96% for different modulated signal when the signal noise ratio is not less than 5 dB.