The recovery of natural gas from unconventional gas reservoirs such as shales and coal beds has increased significantly over the past few years. Since a majority of gas in these reservoirs is in an adsorbed state, knowledge of adsorption behavior of gases is essential for reliable gas-in-place estimates of these reservoirs. Further, some of these reservoirs also offer the potential for carbon dioxide sequestration. The existence of geothermal gradients in such reservoirs can affect the amount of gas. Since the temperatures in a reservoir are generally in the near-critical or supercritical region of the adsorbed gases, an adsorption model should be capable of reliable predictions of temperature dependence of adsorption in this region. To date, limited studies exist in the literature for modeling of temperature dependence of supercritical gas adsorption. In this work, we present a modification of the simplified local density (SLD) model to improve predictions of the temperature dependence of supercritical gas adsorption. Activated carbons were chosen as the reference adsorbents in the development of the modifications to the model since adsorption data on activated carbons are much more plentiful than data on coals and shales, for which data at multiple temperatures are extremely limited. Since activated carbons are simpler structural analogs of other carbonaceous adsorbents such as coals and shales, a model for activated carbons should serve as a basis for extension to the more complex coals and shales. To undertake this study, a database was compiled that includes adsorption data at several temperatures for 11 gases on 18 activated carbons containing about 2600 data points. The database also included adsorption isotherms measured on six coals and six shales at multiple temperatures with a total of 670 data points. The SLD model was modified by introducing a new temperature-dependence expression in the model. The modified SLD model was tested for its efficacy in describing the temperature dependence of near-critical and supercritical gas adsorption on the adsorbents contained in the above database. Results indicate that the modified SW model is effective in modeling supercritical and near-critical gas adsorption. Further, the model was generalized to predict the temperature dependence of adsorption as a function of adsorbate and adsorbent properties. The generalized model was tested with an extended dataset and was found capable of predicting the temperature dependence with an average absolute deviation of 5%, in general. The modified SLD model was tested with adsorption data on coals and shales at multiple temperatures. The model appears capable of describing the temperature dependence of adsorption on coals and shales with reasonable accuracy. We note, however, that the development of a completely generalized model requires additional experimental data on coals and shales and further model testing. (C) 2014 Elsevier B.V. All rights reserved.