Recovering a speech from a noisy environment is the most important work in communication system. The difficulty of this task increased when the signal is attenuated with a babble speech or cross-talk noise. When a mobile user talked in a public place e.g. office, markets, universities, a babble speech noise is created. Then, the corrupted speech needs to enhance in a minimum quality so that the user in opposite side can understand. Speech enhancement techniques are, therefore, developed and improved depending on cost effectiveness and maintaining the minimum quality. This paper contributes to reduce babble speech by developing the dual microphone system (DMS) with two omnidirectional microphones for close talk handset mode in mobile. This thesis compares the quality improvement by using two adaptive algorithms (LMS and NLMS) with different environment. The better position of two microphones is also compared. The DMS system is implemented in MATLAB and the quality improvement is measured with signal to noise ratio (SNR) and SNR improvement (SNRI). The output SNR and SNRI are calculated with different distance between two microphones and with 15 dB and 30 dB attenuation of the signals. The simulation results in this paper show that average SNRI is 15.95 dB with NLMS and 14.71 dB with LMS when the attenuation is 15 dB. Moreover, when the attenuation is 30 dB, then the average SNRI is 16.50 dB with NLMS and 16.56 dB with LMS.