An engineering correlation is presented for the prediction and correlation of the thermal conductivity of pure fluids and mixtures. The model is valid over the thermodynamic states from the dilute gas to the dense liquid and is applicable to nonpolar as well as polar fluids. A key feature of the model is that it provides a practical method for describing the strong critical enhancements of thermal conductivity in the broad vicinity of the vapor-liquid critical point. Quantitative correlations for the thermal conductivity of pure fluids and mixtures require a separation into background contributions and critical enhancements (Sengers, J. V. Int. J. Thermophys. 1985, 6, 203. Sengers, J. V.; Luettmer-Strathmann, J. In Transport Properties of Fluids; Their Correlation, Prediction and Estimation; Cambridge University Press: New York, 1996; p 113). For the background term, we use the model of Mathias and Parekh (Mathias, P. M.; Parekh, V. S. AIChE Annual Meeting, Chicago, IL, Nov 1996), which uses the principle of extended corresponding states, as previously developed by Ely and Hanley (Ely, J. F.; Hanley, H. J. M. Ind. Eng. Chem. Fundam. 1983, 22, 90). We propose a new phenomenological model to describe the strong critical enhancements. The model has been evaluated and validated through data for a variety of pure fluids and mixtures. The critical-point enhancements have been studied and validated for several pure fluids (methane, ethane, carbon dioxide, and R134a) and the methane-ethane mixture. In the noncritical region, the correlation can be used in the predictive as well as correlative modes, but it is recommended that the correlative model be used if data are available to tune the model parameters. In the critical region, component-specific parameters must be adjusted to quantitatively describe the strong enhancement in thermal conductivity.