Abstract: The aim of this study was to quantify the effects of multiple factors on fermentation broth rheology. Industrial fed-batch fermentations of Acremonium chrysogenum were conducted, and rheology properties of samples were adequately described by power law model. Nonlinear modeling taking only fungal morphology and cell concentration into consideration led to poor correlation and little prediction function. One of the reasons probably was that the model was oversimplified and some inconspicuous but significant factors were omitted. In this context, extra elements such as substrate concentration, feed mode, media composition were taken into account, following tremendously increased sample library and existence of variables multicollinearity. Two major morphologies of Acremonium chrysogenum were observed in fermentation broth, i.e., freely dispersed arthrospores and filamentous mycelium. The number of arthrospores was the major factor contributing to rheology properties, based on the standard partial regression coefficients. Using the partial least squares regression (PLSR) model, good prediction of flow index and consistency index can be made from linear recombination of variables, with R²=0.94, R²=0.91 respectively.