Classical Barbiturates are formed by substituting one or both hydrogen atoms at the 5-position with alkyl, aryl, and/or alicyclic groups. In this study, a previously developed UPPER (Unified Physicochemical Property Estimation Relationships) approach is applied to predict the melting points and aqueous solubilities of a series of barbiturates. The descriptors from a previously developed UPPER model on hydrocarbons are used to generate new descriptors for barbiturate ring using multiple linear regression analysis. Melting points can be predicted solely from additive enthalpic and non-additive entropic descriptors. These predicted melting points and aqueous activity coefficients are used to predict the aqueous solubilities. Only three new parameters are added to predict the each of above properties. The average absolute errors in prediction of melting points and aqueous solubilities are 20.6°C and 0.57 respectively. This simple and efficient UPPER approach can be useful for predicting melting points and aqueous solubilities of novel barbiturates and other compounds for which the experimental values are unavailable in the literature.