The xylitol molecule is an important building block that can be used in the production of such interesting chemicals as ethylene glycol and 1,2-propylene glycol. The development of productive processes that enable this transformation depends on various experimental and theoretical information. In order to supply part of this demand, this work sought to study the solubility of xylitol in binary liquid solutions formed by water, ethylene glycol, and 1,2-propylene glycol in the temperature range between 293.15 and 323.15 K, covering the entire molar composition range of the solution. The Jouyban-Acree, NRTL, and UNIQUAC models were used in the correlation of experimental data, and the mUNIFAC model was applied in the prediction of experimental data. In addition, an artificial neural network associated with molecular descriptors was developed to simulate the data. Xylitol showed solubility in the pure components with decreasing values in the following order: water, ethylene glycol, and 1,2-propylene glycol. The solubility in binary solutions had intermediate values according to the intermediate concentration values. The models used proved capable of correlating or predicting the experimental data. The artificial neural networks had a satisfactory performance in the data simulation, and the best observed architecture used four layers of the type 7-3-3-1.