In the enzymatic membrane reactor for separating casein hydrolysatem backflushing technology has been used to decrease the fouling of the membrance. Predication of the backflushing efficiency poses a complex non-linear problem as the system integrates enzymatic hydrolysis, membrane separation and periodic backflushing together. In this paper an alternative artificaial neural network approach is developed to predict the vackflushing efficiency as a function of duration and interval. A contour plot of backflushing performance is presented to model these effects, and the backflushin conditions have been optimized as duration of 10s and interval of 10 min using this neural network, Also, simpleneural networks are established to predict the time evolution of flux before and after backflushing. The results predicted by the modelass are in good agreement with the experimental data, and the average deviations for all te cases are well within +-5% The neural network approach is found to be capable of modeling the backflushing with confidence.