Abstarct: Due to the importance of clixnker grate coolers in cement industry and lack of a mature model for these systems, identification and prediction of the cooler system status are necessary for any simulation and automation approaches. Intrinsically, grate cooler is a MIMO system that in this thesis simulate it using neural network and nonlinear identification method. Since the data used in this research has been extracted from a real system, pre-analysis of data is one of the critical parts of identification process. The results of applying the proposed method on real data of the cement grate cooler is demonstrates the superior performance of this method in performance prediction of this industrial system. Although hardware sensor technology has improved considerably, many variables such as clixnker laxyer thickness or pressure under hot chamber grate are still monitored through laboratory analyses. This is an expensive method and they are inadequate for on-line supervision and considerable delays. To solve this problem, this variable can be estimated by means of measurable variables which this way is so-called soft sensors or virtual sensor. In this thesis utilizes a MLP neural network, soft sensors are designed for those grate cooler variables that are difficult to measure with physical approach. The results show high accuracy sensors is proposed.