The invention discloses a parallel interference cancellation algorithm based on likelihood feedback, and belongs to the technical field of wireless communication. The method comprises the following steps of: (1) receiving a spread spectrum signal by a base station receiver, and modeling the received signal; (2) utilizing a matching filter to carry out modeling on output signals of all users; (3) carrying out modeling on multiple access interference; (4) calculating bit amplitude estimation; and (5) carrying out parallel interference cancellation to obtain a signal after the multiple access interference is cancelled. The algorithm mainly aims at the problems of easiness in signal mutation and poor estimation accuracy in the conventional bit amplitude estimation method, and a novel bit estimation method is provided; secondly, the parallel interference cancellation algorithm based on likelihood feedback is adopted, and a problem of information loss caused by asynchronism when the conventional interference cancellation algorithm is implemented is compensated. According to the algorithm disclosed by the invention, the bit error rate in a transmission process is reduced, and the algorithm has important significance for further application of a wireless communication spread spectrum system.