Abstarct: In this thesis, macerals of input feed to the factory has been detected and a program to mix coals from various zones with different percentages has been presented for efficiency promotion of the factory. To do so, after sampling from input feed of the factory, macerals of all the samples were detected by visual analysis method. By using the design of taguchi experimental method, flotation tests with changing the factors such as collector amount, frother amount, solid content percentage and grain size for all the 5 mines in two levels were done and other parameters in the tests were considered as constant. Results of operational optimum conditions for coals of different mines were determined separately by statistical analysis. Since the coal of these mines is mixed with each other and then is fed to the factory, optimum conditions cannot be obtained without considering optimum compound, so the optimum compound of coal was determined baxsed on Designs Mixture plan. This optimum compound is defined for a specific operational condition. So as a result, by performing signal to noise ratio analysis and considering coal variety as noise factor and control parameters as signal factors, a robust plan with minimum sensitivity to input coal type was presented. The results showed that the most appropriate coal compound by considering the operational conditions and the highest separation efficiency was achieved. baxsed on taguchi design and signal to noise analysis, optimum control conditions are obtained as 2000 g/ton collector, 150 g/ton frother and 20% solid content. By these conditions, it is proposed to treat Takht mine coals sepratly and have Zemestanyurt and Tazare coals blended by equal parts before feeding into flotation cells. The highest sepration efficiency of 88.1% for Takht coal mine and 86.3% for Zemestanyurt and Tazare coal mines obtained.