Having information on the chemical composition of hydrocarbon mixtures (the amount of paraffinic, naphthenic, and aromatics) is still a challenging problem in the area of reservoir fluids. Methods for determining these compositions include experimental techniques and empirical models. Laboratory methods are accurate, but they are costly and time-consuming. Therefore, researchers have tried to use empirical models instead of using laboratory methods. Generally, these models include two or more characteristic parameters such as Refractive Index (RI), normal boiling point (Tb), density in standard conditions (d), carbon to hydrogen ratio (CH), Watson coefficient (K), and viscosity-gravity constant(VGC). The problem with the proposed models is that for finding the compositions the mentioned characteristic parameters must be available, but usually, these parameters are not specified for petroleum cuts, therefore, in practice, the determination of chemical composition is difficult. In this research, we have tried to select parameters for the construction of a model that does not have this limitation. Selected parameters were MW, SG, and Tb which are molecular weight, specific gravity, and normal boiling point respectively. These parameters are commonly available for petroleum cuts, therefore for the determination of the composition of oil cuts, another specific characteristic parameter is not necessary. In this study, the neural network and the SAFT family equation of state have been used to determine the model for estimating the composition of families. For the development of this model, for synthetic cuts consisting of paraffinic, naphthenic, and aromatic, the specific gravity and normal boiling point values were determined using the PC-SAFT state equation. Finally, a neural network model was implemented on these data. In the end, the ability of the model for estimation has been tested using a series of evaluation data for oil cuts. The results indicate that the proposed model predicts the families present in the cuts effectively.