Since sulfur dioxide is released from the oxidation of sulfur in fuel, the level of $SO_2$ in industrial areas can be effectively managed by optimizing the allocation of fuels: the fuel should be allocated to each industries so as to achieve the air quality goal in the area with minimum fuel cost. The solution for this can be by solved using linear programming technique incorporated with the Gaussian dispersion equation. When this method was applied in Ulsan Industrial Complex, 39.3% of fuel expense could be saved compared with the present uniform fuel policy. With this method, bituminous coal generally was allocated to big emission sources such as power plants or to industries sparsely located in remote areas, and LPG or low-sulfur oils to small or medium-size sources in dense indurstrial aras. However, the particulates emission will increase with this policy, because it maximizes the use of coal within the limit to achieve the air quality goal in the area.