The most essential work in a communication system is reconstructing a clean speech from a noisy environment which is a challenging problem in speech processing. The performance of speech processing system could increase by utilizing an effective speech enhancement technique in case of mobile phone, hearing aids, etc. An innovative speech enhancement technique may help in increasing the accuracy of speech recognition methods. In this paper a new approach has been introduced which deals with the implementation of dual channel speech enhancement technique utilizes the coherence function computed from the dual input signals without prior noise statistics for noise reduction. For post-processing, discrete wavelet transform (DWT) is used as a gain function which is set up by SNR values using coherence function. The proposed algorithm is applied in an environment where interfering speakers are present and gives a better result than other algorithms when the speech is corrupted by different noise types. © 2018 IEEE.