The invention provides a multi-access perception source feedback optimization method based on channel learning. The method comprises the following steps: performing feature extraction on a collected sousa chinensis cry signal to obtain prior knowledge; performing quasi-biological signal modelling on the sousa chinensis cry; constructing a marine organism cry smart signal library to output a multi-node quasi-biological signal model in real time; training a signal through a transmission channel, estimating the channel by using a RLS algorithm to acquire a channel impulse response function; respectively making the convolution with a time-reversal function of the channel by the output signal in the smart signal library, carrying the channel information to obtain a fed-back and optimized bionicdetection waveform. The stealth in the underwater environment is improved by using the biological signal; the bionic detection waveform of the channel multi-path information is carried, thereby welladapting to the complex marine channel; the interference of the channel multi-path is inhibited when a receiving side processes the echo, and the more accurate focusing is realized. The method provided by the invention can be well applied to multiple underwater UUV joint operation.