The aim of the present study was to prove that the image analysis is a useful tool in coke texture characterization. The authors proposed the method of measurement of the gray color gradient vector in the prepared microscopic images with visible anisotropic effects. The research was performed for totally different samples of chars and cokes. The total texture fibrosity index for each sample was determined. The results show a strong positive correlation between proposed fibrosity index and technological parameters according to NSC test (CRI, CSR). The new parameter characterizes with universal character resistant to changes in conditions of coking process and coked coal properties. (C) 2014 Elsevier B.V. All rights reserved.