Speech recognition is has been used in wide applications such as Voice-Activated Systems, Virtual Assistants, and Interactive Voice Responses (IVR). Feature extraction is a crucial stage in speech and audio signal processing for various applications, including speech recognition, speaker identification, speaker authentication and audio categorization. The accuracy of any speech recognition technique depends on the type of feature extraction method used. The most commonly used techniques are Linear Predictive Coding (LPC), Mel Frequency Cepstral coefficients (MFCC), and Perceptual Linear Prediction. Among these, MFCC is the most effective and computationally efficient method. This paper focuses on the implementation of MFCC using Simulink for speech signal processing applications. The implementation of feature extraction using MFCC in Simulink involves several steps, including pre-emphasis, framing, windowing, Fast Fourier Transform, Mel-filter bank application, and extraction of cepstral coefficients. Finally, a recurrent neural network is used to detect the presence of the word 'yes' based on the extracted features. In summary, this paper highlights the significance of feature extraction in speech recognition. The Simulink implementation of MFCC feature extraction is explained in detail, and the use of a recurrent neural network for detecting the presence of a specific word is demonstrated.