Aiming at the separation problem of multiple UAV remote control signals, a blind source separation algorithm of multiple frequency hopping signals based on time-frequency analysis is proposed. The difference of the dwell time of FH signals with different hopping periods is used to improve the extraction of the time-frequency ridge.The wavelet transform is used to detect the mutation point of the improved time-frequency ridge, and the maximum dwell time of the ridge is the minimum hopping period of FH signals. The frequency hopping signals with different hopping periods are separated, and the frequency hopping signals with different signal amplitudes are subjected to blind source separation based on different time-frequency energy values. The results show that compared with the similar algorithms, the proposed algorithm can realize the blind source separation of multi-frequency hopping signals in the case of single channel without relying on multi-channel data acquisition and mixing matrix estimation, and has certain engineering application value.