Abstract Human Computer Interaction now focuses more on being able to relate to human emotions. Recognizing human emotions from speech is an area that a lot of research is being done into with the rise of robots and Virtual reality. In this paper, emotion recognition from speech is done in MATLAB. Feature extraction is done based on the pitch and 13 MFCCs of the audio files. Two classification methods are used and compared to determine the one with the highest accuracy for the data set.