Abstract: A fast adaptive algorithm for noise cancellation using multi-sensory signal recordings of the same noise source was described. In order to avoid the trap of local minima, a design method based on the genetic algorithm was introduced to treat the nonlinear optimization of the fixed-order optimal adaptive noise cancellation. The designed method utilized genetic algorithm to adjust the weights of linear combiner(LC) of multi-sensor signals adaptively and reduced the large estimated error of the averaging method of all sensor signals. The design method can obtain simplicity of implementation, conservation of operation time and better reconstruction performance by fully utilizing multi-sensor signals.