Aiming at exploring the possibility of building a model by artificial neural network for the multiple-effect evaporation in sugar factory, the error BP（BackProragation）neural network is used to establish a model and describe the complex relationship among the juice steam pressure P1、P2 of the first and the second stage and the syrup brix C5 leaving the multiple-effect evaporator. And the model is implemented by Matlab7.0. The simulation result proves the validity of the model, and it aslo lays a foundation for the further investigation into the intelligent control of the multiple-effect evaporation system.