It is well known that the communication in a noisy environment is a difficult problem to solve for both normal and hearing impaired listeners. In this work, the signals obtained from the omnidirectional microphone which is mounted just above one ear on the head of the KEMAR mannikan and a hypercardioid microphone which is placed at the center of the vertebral column on the back were used as input to the two channel adaptive noise canceller. The intelligibility tests conducted by the authors are presented in this paper to illustrate the performance of both Linear and Non Linear model. Results show that the Volterra LMS scheme is very effective in improving intelligibility and SNR for different environmental noise.