An artificial neural network with a culture database was developed to predict process behavior in cheese production. Based on the experimental investigations, it was found that the determination of the final process time of the acidification step, followed by rennet addition, is the key to successful operation of cheese processing. In order to determine the optimal timing for the rennet addition, it is practically useful to develop a model which can predict the final process time in the acidification process for successful cheese production. Therefore, an artificial neural network system with a culture database containing various operating data, in which the learning data for back propagation were selected from a culture database based on the Euclid distance, was examined. The system enabled successful prediction the final process time in the acidification process based on several operating data and the culture database. (C) 2003 Elsevier Ltd. All rights reserved.