The trained versions of Jouyban–Acree and modified Wilson models are proposed to predict the drug's solubility in {acetonitrile + water} solutions at different temperatures. To obtain a full predictive model, the Abraham solvation and Hansen solubility parameters of solutes representing the interaction between components are combined with the proposed models. The solubility data of 12 drugs are fitted by these models. The overall average relative deviation percent (%ARD) values for the back-calculated solubility of drugs in aqueous acetonitrile solutions are 18.3–52.3 for the proposed models. The generally trained models provide acceptable estimation of the solubility of drugs and can be helpful in the pharmaceutical sciences.