The embodiment of the invention discloses a method and a device for enhancing voice. One specific implementation mode of the method comprises the following steps: acquiring time domain voice, acquiredby a microphone array, of a plurality of channels; based on the time domain voice of the plurality of channels, generating frequency domain voice of at least one channel; analyzing the frequency domain voice of at least one channel to acquire normalization enhanced coefficient of the frequency domain voice of at least one channel; performing enhancement processing on the frequency domain voice ofat least one channel by utilizing the normalization enhanced coefficient of the frequency domain voice of at least one channel to acquire enhanced frequency domain voice of at least one channel; performing inverse Fourier transformation on the enhanced frequency domain voice of at least one channel to acquire enhanced time domain voice of at least one channel. According to the implementation mode, targeted voice enhancement is realized, noise and room reverberation in the voice can be better eliminated, and the accuracy degree of voice identification is increased.