The industry of long products in hot rolling mills are slowly transforming into a datadriven business. Large amounts of data are being collected at mills around the worldin hopes of improving production quality and saving costs. This report examines thisdata in order to improve process simulations of rolling mill products. Data processing was conducted to identify key data points. The data points were compared withrolling simulations using Wicon. Modelling parameters, such as spread coefficients,was then optimized with objective to minimize the residual between the simulation and the measured data from the mill. The optimization was conducted using two different optimization methods (Levenberg-Marquardt least-squares and Multi-objective Differential Evolution). More accurate simulation parameters can improve the process engineers ability to construct rolling schedules with lower tolerances in production. The results show that better simulations with unique optimization parameters for different products is possible. However, more data must be processed in order tovalidate the method.