The application of the extended corresponding states modeling technique to thermodynamic and transport properties has demonstrated that different conformality behaviors are followed by a sarne group of fluids inside each of these two categories of properties. The traditional extended corresponding states technique for transport properties requires the conventional thermodynamic shape factors, derived from accurate equations of state for both the target and the reference fluid, and an additional shape factor, derived from transport property data of the target fluid, in order to fit the transport properties themselves. A new extended corresponding states model is here proposed for thermal conductivity; the technique uses two shape factors, one for each independent variable, which are generated just from the available thermal conductivity data in the range of interest. As a consequence there is no need to import shape factors from thermodynamics. The shape factors are obtained as continuous functions through a neural network, because of its flexibility and high capability to fit the data. The accuracy of thermal conductivity data representation on the basis of this model is better than that achieved through conventional approach by summation of dilute gas, excess and critical enhancement contributions. (c) 2006 Elsevier B.V. All rights reserved.