The production of Fructooligosaccharides (FOS) was carried out by applying two stage continuous process. In the first stage Fructosyl Transferase (FTase) from Aspergillus oryzae CFR 202 was grown in sucrose containing medium. In the second stage the enzyme was used to produce FOS with sucrose as the substrate. Though the processes were two stage, the system was continuous as the enzyme prepared was immediately used to prepare the FOS. Yields of FOS production were maximized using response surface methodology (RSM) based on shell design. RSM was used to evaluate the important parameters that influence the production of FOS in both the stages. With this the feasibility of developing a maximization programme for a continuous two stage process was demonstrated. Fermentation time (36-108 h). KH2PO4 concentration (0.2-1%) and sucrose concentration (1 24%) in the fermentation medium, reaction time (1-24 h) and pH of the reaction mixture (5-6) were chosen as process variables for the optimization. Among these parameters, reaction time and fermentation time had significant effects compared with KH2PO4 concentration. sucrose concentration and pH of the reaction mixture. Optimum conditions for the production of maximum FOS yields were fermentation time-108 h. KH2PO4 - 0.723%. sucrose-6.455%. reaction time - 18 h and pH of the reaction mixture 5.15. The coefficient of determination (R-2) for the model was 0.97. Repression probability (P) was 6.98 x 10-(16). The maximum FOS yield predicted by the equation (58.9%. w/w) agreed well with the values obtained from the experimental verification (56-4%. w/w) at the optimum values based on stationary points. To have few more options for higher yields of above 50%. contour plots were also used to predict the experimental conditions. This maximized the FOS yields at 58% (w/w). These optimum conditions were then scaled up to 10 1 level of FOS production and the results were matching the shake flask level studies. (C) 2004 Elsevier Ltd. All rights reserved.