A model based on solubility parameters is proposed to predict the solubility curves of antihelmintic drugs at several temperatures, including aqueous and non-aqueous mixtures. The solubility of the drugs was measured in ethanol-water and ethanol-ethyl acetate mixtures at 15-35 degrees C (mebendazole) and at 25 degrees C (thiabendazole and metronidazole). The solid phases were analyzed by differential scanning calorimerty. The polymorphic form A of mebendazole was also characterized from infrared spectroscopy. Markedly different solubility profile shapes were obtained against the solubility parameter of the mixtures: two symmetrical peaks (metronidazole), two maxima of different height (mebendazole) and a single peak (thiabendazole). The solubility parameter of the drugs was related to the co-solvent action of both mixtures and to the solubility peaks. The single equation proposed was able to predict solubility profiles of different shape, including both mixtures and all temperatures, providing reasonable physical meaning for the regression coefficients. The model was successfully tested for its predictive capability using a limited number of experimental data. More than 100 solubilities were predicted at several temperatures using 20 data point for each drug.