Abstarct: In this work, a quantitative structure-property relationship (QSPR) approach was employed to predict the density of organic liquid mixtures over a wide range of temperatures and pressures. Two types of descxriptors for mixture components have been used for calculating of mixture descxriptors: group contribution baxsed descxriptors and theoretically derived descxriptors. In the first type, the descxriptors of mixture components are selected baxsed on group contribution method (GCM). After analyzing the chemical structure of all studied mixtures component in this work, 15 functional groups were recognized. In the second method, the 1481 descxriptors of individual components of the mixture were calculated using dragon software. After calculating of descxriptors for mixture components with two methods, the mixture descxriptors were calculated as mole-weighted sums using the descxriptor value and mole fraction of each pure component in the mixture. Therefore, 15 descxriptors were obtained by GCM for mixtures. However, the best calculated descxriptors from a large number of theoretically derived descxriptors of mixtures may be selected by genetic algorithm baxsed on partial least square (GA-PLS) method. 22 theoretically derived descxriptors have been selected by GA-PLS method. The selected descxriptors with two feature selection methods and 2 experimental variables (temperature and pressure) as well as total mass of mixture were used as input nodes for constructing different artificial neural network (ANNs). 4 feed forward artificial neural networks were designed by combination of Levenberg-Marquardt (LM) and Bayesian regularized (BR) algorithms with logarithm-sigmoid and tangent-sigmoid transfer functions. After training and optimization of the ANNs parameters such as, number of neurons in input and hidden laxyers, and number of iteration (epoch), the performance of the models was investigated by the test set. The results showed that the Bayesian regularized artificial neural network with tangent-sigmoid transfer function and GCM baxsed descxriptors, may be simulated the relationship between the structural descxriptors and density of the desired mixtures accurately. The mean square error (MSE) and determination coefficient (R2) for the test set are 11.051 and 0.9996 respectively.