The development of continuous predictive correlations for the temperature dependence of the ideal gas heat capacity for hydrocarbons and substituted hydrocarbons valid over a wide range of elemental compositions (CHNOS), molecular structures, and temperatures (298.15-1500 K) is reported. The correlations are functions of temperature, elemental composition, and 10 (naphthenic hydrocarbons) or 12 (aromatic and aliphatic hydrocarbons) universal coefficients. There are no compound specific coefficients. The variables are absolute temperature and a similarity variable, which possesses a value proportional to the number of atoms in a molecule, irrespective of their nature, divided by molecular mass. Different data sets obtained from the NIST TRC Ideal Gas Database 88 were used to evaluate parameters, and to test the predictive nature of the correlations. Detailed comparisons between predicted ideal gas heat capacities (c(p)(g0)) obtained in this work and values obtained using four Lee-Kesler correlations, and the Harrison-Seaton correlation are presented. The present correlations are shown to be preferred over these other options on the basis of accuracy and range of application for general-purpose calculations, unless the fluid assignment (naphthenic vs. aromatic/aliphatic) is unknown and c(p)(g0) values are only required at set temperatures available for the Harrison-Seaton correlation. In this latter case, the Harrison-Seaton correlation is preferred. For large molecules, the universal correlations presented in this work approach the accuracy of the benchmark Benson method. The universal correlations are well suited for the determination of ideal gas heat capacities of large and complex molecules or mixtures, where more precise structure-property group contribution correlations are difficult to apply or are inapplicable because groups present are not defined in the correlations or are unknown. (C) 2013 Elsevier B.V. All rights reserved.