The main production method of branched chain amino acid (BCAA) is microbial fermentation. In this paper, to monitor
and to control the fermentation process of BCAA, especially its logarithmic phase, parameters such as the color of
fermentation broth, culture temperature, pH, revolution, dissolved oxygen, airflow rate, pressure, optical density, and
residual glucose, are measured and/or controlled and/or adjusted. The color of fermentation broth is measured using the
HIS color model and a BP neural network. The network's input is the histograms of hue H and saturation S, and output is
the color description. Fermentation process parameters are adjusted using fuzzy reasoning, which is performed by
inference rules. According to the practical situation of BCAA fermentation process, all parameters are divided into four
grades, and different fuzzy rules are established.