The ability to correlate and predict the solubility of solids in supercritical fluids is of the utmost importance for the design and the evaluation of supercritical processes. Previously, we have investigated the solubility of a pharmaceutically interesting solid compound in supercritical carbon dioxide, alone or mixed with cosolvents. In this work, these solubility data are correlated through several density-based semi-empirical models. These models have been either modified or extended to be applied to mixtures including a cosolvent. The validity of the resulting correlations is checked by using the solubility data of another pharmaceutical solid, naproxen. (C) 2003 Elsevier Science B.V. All rights reserved.