In cement manufacturing, the important process is the process to proportion raw materials.A good raw mixed material is very difficult to produce because of the fluctuationof chemical compositions of raw materials. The quality of a raw mixed material is determined by three values: Lime Saturation Factor (LSF), Silica Modulus (SM) and Alumina Modulus (AM).Although, equations defined these three qualities are linear equations, a simple linear programming can not be used to determine the practical values for these three qualitiesdue to the limitation of feeder capacity. In other words, the proportion of each raw materialcan not be any values. Therefore, we design a method using genetic algorithm to control the proportion of raw materials while having a quality of raw mixed material in an acceptablerange. The genetic algorithm actually defines the optimal proportions of raw materialsat certain period of time by minimizing the deviation of the actual values from the expectedvalues of LSF, SM, AM and total production rate and selecting the best proportionof raw materials for the next cycle. We evaluated our method by using computersimulation using all setting as ones used in the real cement manufacturing. From the simulation results, our method can determine the proportion of raw materialhaving a quality in an acceptable range although the chemical compositions of each raw materialfluctuate from time to time.