Optimization of operating parameters for achieving maximum yield of cleans at the desired ash level in composite coal washeries is a major problem, particularly when the washeries receive raw coals from a large number of mines and there is a frequent variation in the proportion of coals from different sources or in their individual characteristics. This problem is more acute in countries with a predominant reserve of difficult-to-wash coals. Control of parameters in such cases requires too frequent sampling and characterization of the blended feed as the graphical method for parameter optimization based on individual coal characterization is not practicable for large numbers of source coals or washing size fractions. This paper presents a solution in the form of a numerical method with any number of source coals or washing size fractions, with the help of constrained non-linear optimization of a multivariable objective function. Operational constraints of different washers in the composite plant functioning on the principle of gravity separation can also be largely overcome as any variation in the elementary ash is incorporated in the method in the form of a set of inequality constraints along with the range of (available) specific gravity of cuts and desired ash content of the overall cleans. A case study with nine coals of different characteristics under two different mode of operations is presented.