At the first step of the present study, the modeling of the conventional methanol production unit is performed and its results have been validated using three-day industrial plant data. In the second step, the great effort has been made to improve the performance of the conventional unit. Sensitivity analyses confirm that decreasing the temperature and increasing the pressure, can move the process towards increasing the amount of methanol production as well as reducing the byproduct. Hence in the proposed modified scheme, the effluent of the carbon dioxide hydrogenation reactor is cooled and then sent into a new added reactor. The performance of the conventional and modified methanol production units have been modeled using artificial neural network. The obtained results show that the amount of methanol production by modified scheme is two times more than conventional one.