From 61 coking coals, 36 coal blends were prepared. Using a pilot coke oven, cokes were prepared from both 61 coking coals (Type I cokes) and 36 coal blends (Type II cokes). Coals were characterized by 14 coal characteristics and cokes by Coke Reactivity Index CRI and Coke Strength after Reaction with CO2 CSR. For the study of mutual statistic relationships among experimentally determined characteristics of coals and cokes, the Factor (FA) and Regression Analyses (RA) were used. FA distributed characteristics of coals and Type I cokes into 4 factors while characteristics of coal blends and Type II cokes were distributed into 7 factors. In case of pure coals and Type I cokes, strong relationships with high correlation coefficients (R > 0.60 ) were more abundant than in case of coal blends and Type II cokes. FA was used for the selection of coal characteristics that influence the coke quality the most significantly. These characteristics were then recalculated by RA for the predictions of CRI/CSR of Type I cokes. Predictions of CRI/CSR of Type II cokes were calculated from coal blends by the same procedure. The comparison of the predicted and experimentally determined CRI and CSR indexes showed much more reliable prediction of CRI/CSR indexes calculated from coals than calculated from coal blends. This study also explains the dominant reasons of this observation., Jana Serenčíšová, Zdeněk Klika, Ivan Kolomazník, Lucie Bartoňová and Pavel Baran., and Obsahuje bibliografii