Coal raw material base of coking is the main factor characterizing the quality of coke. Therefore, it is very important to know technological properties and peculiarities of coals behavior in a charge during coking process for coals charge batching and coke quality control. One of the priority directions in study coals and charges is petrographic and reflectogram analysis, which enable to obtain data related to evaluation genuine (one-valued) technological properties of coals, coal blends and charges at production of coke of required quality. Using a broad material of study, including the one carried out by the authors of the article, a wide range of application of reflectogram analysis of coals, coals blends and charges in the coking production was shown. It was demonstrated also that application of the analysis enabled to exclude the problem of “twins”, to define the degree of genetic coals recoverability and coals grades or types relation in the mixtures for the coking. Based on the elaborated by the authors reflectogram criteria of charges for coking, a strategy of coals batching was proposed, which ensures production of metallurgical coke of required and high quality and safe running of coke ovens. Based on wide experimental studies of plastic-tough properties of coal charges, porosity of coke, its X-ray structure characteristics, strength and reaction ability, theoretical ideas were formed about mechanism of interaction in a charge of petrographically nonuniform coals comprising it during coking process, by using the proposed by the authors indices of coals nonuniformity. The indices of coals, comprising the charge, nonuniformity, differ by metamorphism degree (σR) and petrographic composition (σСК), explaining regularities of forming of quality of coke from the charge with participation of petrographically nonuniform coals. The package of the factors noted by the authors, revealed in the process of the study of coals, coals blends and charges, as well as quality of coke obtained from them, enabled to elaborate a complex index of charges coking ability (К.п.к.Vo), which enables to considerably simplify the mathematical model of coke quality prediction and to increase its reliability. Mathematical models of coke quality prediction were verified and implemented at several plants of Russia.