On most Internet software or applications, speech is still a vital means of human communication. When the speech is received and understood by machines, however, ambient noise such as machine noise or AWGN (Additive White Gaussian Noise) is frequently introduced, which has a negative impact on the quality, clarity, and intelligence of speech communication. Therefore, the noisy speech here needs to be processed with speech enhancement technology so that the speech quality, clarity, and intelligence are expected to be improved. This project employs the modified spectral subtraction in speech enhancement technology to process the MATLAB-tested noisy speech based on the aforementioned issue. The modified spectral subtraction includes an over-subtraction factor and a compensation factor based on the concept of over-subtraction, as well as two modes based on amplitude spectral subtraction and power spectral subtraction. The findings indicate that the modified spectral subtraction could enhance noisy speech to some degree.