Emotions are the feelings of a person and the reaction to a situation. People can verbally or nonverbally convey their feelings. In the literature ofSpeech emotion recognition, many of the techniques have used traditional ways to detect emotions, this paper present different way of recognizing emotions, by speech signals are processed by CNN to extract features, the extracted features are then used to input SVM. The SVM outputs the predicted emotions. From this approach we improved accuracy by testing and training the models based on the input audio data