An ingenious device in the use of a multiple regression analysis model for better prediction of state variable changes in bio-processes is proposed. A correction factor was introduced in the regression equations for the estimation of specific rate parameters taking into account the data distribution resulting in the improvement of the prediction of glucose isomerase production by a Streptomyces fusant during a batch cultivation with periodical monitoring.