The power plant owner is interested to know in advance the quality of coal to be burnt which should meet maximal efficiency of power plant and the environmental regulations. There is the need to control and to predict the quality of coal at the mine site to meet sulfur emission requirements. Coal quality control between the mine site and the utility plant is a complex problem owing to the variable nature of coal properties (heating value, sulfur, ash), even within the same coal seam. Due to the fluctuation of the coal quality, mine planning and coal homogenization are in fact an optimization problem under uncertain conditions. To deal with these issues a stochastic optimization model is developed for an illustrative coal homogenization problem. Mining block grades from an optimized mining schedule are used to simulate any given homogenization process in stockpiles throughout the mine`s life. Sulfur content is treated as lognormally distributed random variable. The objectives of the model include minimizing the expected sulfur content and standard deviation in sulfur content. The methodology was illustrated using the case study on Kolubara surface coal mine.