The invention belongs to the technical field of radio-frequency spectrum sensing, and particularly relates to a parallel cooperation spectrum sensing method based on the genetic algorithm. The parallel cooperation spectrum sensing method includes the steps of generating an initial group, conducting individual amendment, calculating a fitness function, selecting an individual with the highest adaptability, conducting variation, struggling for existence, conducting winner intersecting, generating individuals at random, conducting individual amendment, enabling the most adaptable individual to directly enter a next generation, and stopping. By means of the parallel cooperation spectrum sensing method, the differences of different sensing users on different kinds of heterogeneous channel sensing performance, the system effective throughput and the sensing time can be sufficiently considered, and the optimal sensing strategy, namely the sensing task distribution and the sensing time length, is determined. Compared with a random sensing method, the larger system effective throughput can be obtained.