A novel method for predicting the Gieseler maximum fluidity (F-max) of a coal + plastic mixture formed from a relative proportion of the plastics present in a multicomponent waste is proposed. A training set of five most-common thermoplastics in household wastes (HDPE, LDPE, PP, PS and PET), binary and ternary plastic mixtures was used to construct multivariable linear regression (MLR) models. Validation was conducted by means of an external set of mixed plastics and real unsorted plastic wastes. The results obtained from the numerical solution of the MLR models were found to be in satisfactory agreement with the experimental data obtained using a Gieseler plastometer. The F-max values fitted the models with determination coefficients of >0.96 and root mean square errors of prediction of 0.048 and 0.058. All the plastic mixtures tested represented a wide spectrum in concentration of the five polymers contained in municipal plastic wastes and a global plastic addition of 2 wt% to the coal was always used. The starting point for this study was to determine the effect of each single plastic on the reduction in fluidity of various coking coals and an industrial coking blend. Afterwards, the exponential functions of Fmax of the blends of coal and binary/ternary plastic mixtures were useful to analyze the changes in Gieseler Fmax with varying proportions of components. Based on the results, the coal responses were statistically treated and MLR models were developed. (C) 2015 Elsevier Ltd. All rights reserved.