An alternative way is presented to evaluate the possibility of supercritical fluid extraction of a solute when no experimental data on solubility is available. To predict the solubility Of L-Dopa (a drug commonly used for the treatment of Parkinson's disease) in supercritical CO2, the Peng-Robinson equation of state is applied in a fully predictive way with UNIFAC group contribution activity coefficient model incorporated in its mixing rules using the linear combination of Vidal and Michelsen (LCVM) procedure. Experimental solubility data of caffeic acid in SC-CO2 is used as the reference basis for this methodology and the results for L-Dopa solubility suggest further experimental studies. (c) 2004 Elsevier B.V. All rights reserved.