Abstract: The potential use of artificial neural networks to describe liquid–liquid phase equilibria in ternary systems under polythermal conditions is considered. The study was carried out on the example of ten ternary systems, including binary splitting subsystems of water–esters of carboxylic acids, which determines the phase splitting in ternary systems (the third component is alcohol or carboxylic acid). The features of the selected network architecture are presented, and the results, with a critical assessment of the accuracy of the calculations, are given in the tables. Approximations based on artificial neural networks are compared with calculations based on the non-random two-liquid (NRTL) model.