Aims: Models for predicting solubility of drugs in solvent mixtures have an important practical application. Solubility behavior of pioglitazone hydrochloride in solvent blends ranging from non-polar to highly polar is essential. So the present investigation deals with study of pioglitazone hydrochloride in binary solvent systems. Methods: The solubility of pioglitazone hydrochloride in various dimethylsulfoxide-water mixtures was analyzed in terms of solute-solvent interactions using modified Hildebrand-Scatchard treatment. The solubility of pioglitazone hydrochloride in dimethylsulfoxide-water shows a curve with solubility maxima well above the ideal solubility of the drug. Results: The discrepancy between the results using the original Hildebrand-Scatchard equation and experimental points demonstrates that regular solution theory cannot be used to predict drug solubility in dimethylsulfoxide-water binary systems. This behavior has been dealt with the theoretical replacement of mean geometric solubility parameters (δ1δ2) term with the interaction energy term (W). This is attributed to solvation of drug with the dimethylsulfoxide-water mixture, and indicates that the solute-solvent interaction energy (W) is larger than the geometric mean (δ1δ2).The new approach provides an accurate prediction of solubility once the interaction energy ‘W’ is obtained. In this case, the energy term is regressed against a polynomial in δ1 of the binary solvent mixture. Quadratic, cubic, and quartic expressions were utilized for predicting the solubility of pioglitazone hydrochloride in dimethylsulfoxide-water mixtures. But a quartic expression of ‘W’ in terms of solvent solubility parameter was found appropriate for predicting the mole fraction solubility and yields an error ~27.68%, a value approximating that of the experimentally determined solubility. Conclusions: Extended Hildebrand Solubility Approach was successfully applied to reproduce the solubility behavior of pioglitazone hydrochloride in dimethylsulfoxide-water binary mixtures within the accuracy. The method has potential usefulness in preformulation and formulation studies during which solubility prediction is important for drug design.