In this work, a model was proposed for predicting heat capacity of gas (Cp) for diverse organic compounds including aromatic, acyclic aliphatic hydrocarbons, ketones and acetates at variable temperatures, range from 50 to 1500 K. At the same time, a new norm descriptor based on molecular characteristic matrix was developed. Validation results showed that this norm-index-based model had satisfactory performance for predicting Cp at various temperatures with high squared correlation coefficient R-2 of 0.9976 and high squared relation coefficient of the cross validation Q(2) of 0.9976. The results of external validation, R(2)training and Retest, were 0.9977 and 0.9970, respectively. Meanwhile, Y-randomization test and mean absolute errors test could indicate the stability and accuracy of this model. In addition, comparison results could further demonstrate the high predictive performance and the wide applicability domain of this work; and the big strength of this work is that it could predict gas heat capacity down to 50 K. (C) 2017 Elsevier B.V. All rights reserved.