Aiming at the problem that it is difficult to detect the key biological parameters in the process of microbial fermentation

a real-time learning modeling method is proposed. Firstly

considering the multi-stage characteristics of microbial fermentation

fuzzy C-means（ FCM） clustering algorithm is used to cluster the fermentation data samples. Then

local weighted partial least squares（ LWPLS） with just in time learning strategy is adopted as the local modeling method

because its bandwidth parameter have an important influence on the predictive performance of the model. In order to further improve the predictive precision of the model

using the particle swarm optimization（ PSO） algorithm to optimize bandwidth parameter. Finally

the alkaline protease fermentation process is taken as an example. The experimental simulation results show that the proposed the FCMPSO-LWPLS soft sensor modeling method has high prediction precision

it can adapt to real-time monitor of key biological parameters（ total sugar concentration S） in general fermentation process.