: Key tools in the rolling processes are the rolls themselves.They have to withstand severe extremes of temperature and load.The wear of work and backup rolls has the major influence on final product quality by changing their geometry and surface roughness.The principal aim of this thesis is focused on the relationship between the rolling program and pass schedule with work roll wear profile. This has been presented with the neural network method in order to optimize the rolling program and the schedule for the axial roll shifting.Furthermore, an optimization process of rolling program with Genetic Algorithm (GA) has been carried out. Due to an abnormal condition existing on industrial nature of the factory, the process of moving from one pass to another and also due to the complexity of tribological systems and simultaneous presence of different mechanism which causing the wear, the phenomenological approach for predicting of wear are still imperfect and quite complicated for a practical use. Main restrictions which occur here are including the number of the included influence parameters or variables, their weights, and the consideration of their mutual space interactions. Neural network with having enough weights and biases can understand the latent effective parameters on the work roll wear behavior for prediction of wear profile. Some of the phenomena that happen on practical mill are as follows; spreading of plate, shrinkage on final stands because of decreasing temperature and the creep between stands because of high temperature in which makes this phenomena so difficult to surgery of hot roll.In this order, the length of rolls has been divided to distinct segments in which for each segment, a separately neural network is defined. For making unit neural network for all of the segments with similar input parameters and different outputs parameters, particular neural network is defined. This action saves some time on the modeling. At the end, this particular neural network is used for preparing a target function on the base of effective parameters for optimization of rolling program by GA. Preparing of optimal rolling program with minimizing wear and also uniform wear profile can prevent production of no quality sheet metal which is important on economical and quality aspects of production.This work refers to the practical data from SABA Rolling Strip Complex. Neural network could truly predict work roll wear profiles for new rolls and also predict some phenomena such as local wear due to high pressure on plate edge. Algorithm genetic could prepare rolling program with minimizing wear and uniform wear constraints on roll length. Key words: Hot strip mill, Wear profile, Artificial Neural Network, Genetic Algorithm, Rolling Program, Optimization