Abstract: In order to reduce the influence of interferences and noise on frequency hopping signal detection under low signal-to-noise ratio, a blind detection algorithm for multi-frequency hopping signals based on time-frequency analysis is proposed. Considering the different time-frequency distribution characteristics of frequency-hopping signals, fixed-frequency signals, and Gaussian white noise, the time-frequency cancellation ratio is constructed by using the time-frequency map which is acquired from short-time Fourier transform. It is used as the detection statistics because the value of the time-frequency cancellation ratio of each signal is different by the theoretical analysis, and the blind detection of frequency hopping and fixed frequency signals under the background of Gaussian white noise is achieved. The simulation results shown that this algorithm has good robustness to noise power uncertainty, and compared with the improved power spectrum cancellation method, this algorithm has a higher detection probability of multi-frequency hopping and fixed frequency signals in a low SNR environment. This algorithm also can achieve the blind detection of multi-frequency hopping signals in a complex electromagnetic environment where the fixed-frequency signals、sweep signals and burst signals exist as jamming signals, it can improve the signal-to-noise ratio by 10 dB compared with the compared method, when the signal-to-interference ratio is 5 dB and the detection probability of frequency-hopping signals reaches 100%;and the false alarm probability is close to 0 when the interference-to-noise ratio is 0.05 dB.