This paper considers classification of human body odor based on the neural network of learning vector quantization (LVQ). Odors of human body are sweaty odor, middle-aged odor, and age-of-odor, which are called three body odor. The sweaty odor is detected for humans from teenagers to twenties. the middle-aged odor is for humans from thirties to fifties, and the age-of-odor is for humans over fifties. We classify odors into three odors stated above. The sweaty odor is an odor similar to ammonia and isovaleric acid, middle-aged odor is similar to diacetyl, and the age-of-odor is similar to 2-nonenaar. Utilizing a special sampling box, we measure the odors first and classify the chemical products such that such that each of those odors could be classified into true odor using LVQ. After that, we measure the real human body odors for two thousand and one hundered persons. Finally, we develop a hardware (Kunkun body) to classify various odors into each odor among three odors.