We propose a new method of adaptive microphone array-based speech enhancement. This approach is based on the enhancement of SNR(signal to noise ratio) in the speech signal and improvement of speech recognition performance using the delay-and-sum beamforming method with the LMS(least mean square) algorithm. In the performance evaluation, the microphone array-based speech enhancement using the DSBF(delay-and-sum beamforming) with the LMS algorithm shows about 5 [dB] of SNR enhancement and 21 percent of error reduction in the spontaneous speech recognition domain.