As of late, Automatic speech recognition has advanced on account of instruments, for example, natural language processing, and deep learning, among others. It is a framework or put in another way, a gadget that changes a raw signal into computer comprehensible text. The genuine creation of speech is comprised of changes in air pressure that outcomes in pressure wave that our ear and cerebrum comprehend. The vocal tract is utilized to deliver a human speech, which is adjusted by teeth, tongue, and lips. Speech recognition alludes to a machine's ability to perceive human speech and transform it into a computer comprehensible text. Speech recognition is a magnificent illustration of good interaction between humans and computers. In this paper, we introduce the process to extricate the feature from the signal utilizing Mel-frequency cepstral coefficients. Mel-frequency cepstral coefficients are a genuinely far wide and proficient methodology for feature extraction from a sound file. This technique improved the speech recognition process and removes the distortion in the voice. In this manuscript we applied the Mel-frequency filtration process to improve speech and remove the background noise. the Therefore, the proposed methodology gives better performance in the automated speech recognition system.