As we all know, Emotionalization will not only have a negative impact on people's body and mind, but also make people make mistakes in judgment. If there is a portable "management system", which based on human speech. And then, through the mathematical parameter extraction of its speech signal, the formant, MFCC and other parameters are obtained. Finally, it is compared with the sample signal of speech database and classified by KNN algorithm, and the emotion classification of the test speech signal is obtained. At the same time, we also designed a system framework for carrying voice input, emotion detection, cloud storage and other functions to realize a complete speech recognition and detection process.