A systematic approach is suggested for predicting the solubility of sparingly soluble solid fine chemicals and pharmaceuticals. The procedure uses group contribution methods for computing the difference in solubility at infinite dilution in the solvent of interest from an optimal reference solvent with the aim of (1) minimizing the impact of uncertainties in pure-solute properties, (2) decreasing the number of adjustable parameters to be determined by data reduction, and (3) using appropriate experimental data to fit unknown parameters. Several examples illustrate the method.