A thermal power plant boiler is a non linear process in which steam temperature changes in a non linear fashion with change in load. Other parameters of boiler that change with change in load are steam pressure and steam flow. The changes in above said parameters affect the overall efficiency of a thermal power plant. To match the load requirements it is necessary to control these parameters. In this paper, steam temperature of the boiler is controlled by simulation under MATLAB simulink. The steam temperature of the boiler is depends on the heat transfer rate inside the boiler which is directly proportional to the fuel flow rate. Because of the multi input multi output, non-linear modeling capabilities of a neural network, it is a powerful alternative for modeling of complex process of heat transfer inside a power plant boiler. The trained neural network is used to adjust the fuel flow rate for minimizing the effect of firing system disturbances on steam temperature.