In this work, a correlation for the estimation of the theta temperature (lower critical solution temperature, LCST) in polymer solutions is proposed in which both polymer and solvent molecules are characterized by the molecular connectivity index. A database of 169 data points, including 12 polymers and 69 solvents, was adopted to validate the new correlation. The results show that the correlation gives good estimations of theta(LCST), with an r(2) value of 0.77 for the training set of 112 systems and an r(2) value of 0.75 for the test set of 57 systems. Because one correlation was applied to all kinds of polymer solutions and the correlation requires only the molecular connectivity indices of the solvents and polymers in the calculations, the new correlation is a generalized predictive model that is easier to apply and has better predictive capability than the existing methods and models.