The relationship between maceral content plus mineral matter and gross calorific value (GCV) for a wide range of West Virginia coal samples (from 6518 to 15330 BTU/lb; 15.16 to 35.66 MJ/kg) has been investigated by multivariable regression and adaptive neuro-fuzzy inference system (ANFIS). The stepwise least square mathematical method comparison between liptinite, vitrinite, plus mineral matter as input data sets with measured GCV reported a nonlinear correlation coefficient (R 2) of 0.83. Using the same data set the correlation between the predicted GCV from the ANFIS model and the actual GCV reported a R 2 value of 0.96. It was determined that the GCV-based prediction methods, as used in this article, can provide a reasonable estimation of GCV.