Accurate estimation of two-phase compressibility factor (Z factor) in gas condensate reservoirs is essential for understanding phase behaviour and reliable simulation studies. Experimental measurements of Z factor in wide operational conditions are cumbersome and costly. Hence, engineers rely on existing models such as cubic equation of states (EoSs) and empirical models for estimating such important property. In this study initially the accuracy of prevalent available Z factor models for prediction of two-phase gas condensate Z factor were examined. Then, several smart models including two multilayer perceptron neural networks known as feedforward neural network (FFNN) and cascade forward neural network (CFNN) optimized with Levenberg-Marquardt (LM) and Bayesian-Regularization (BR) algorithms and one Adaptive Neuro Fuzzy Inference System (ANFIS) optimized with Particle Swarm Optimization (PSO) were developed based on 19518 data points for the same task. The databank covers gas condensate two-phase Z factor, compositional variations, molecular weight of heptane plus (MWC7+) and gas specific gravity in wide range of pressure and temperature. The results indicate that the FFNN-BR predicts all experimental data with high accuracy with an average absolute relative deviation of 0.321%. Furthermore, the accuracy of the developed model over two cubic EoSs and three empirical models from literature was confirmed. Finally, based on the sensitivity analysis, it was found that the pseudoreduced pressure (Ppr), MWC7+, and molar content of C7+ in the mixture have the highest impact on prediction of two-phase Z factor. The proposed tools can be utilized for accurate prediction of gas condensate two-phase Z factor to ensure accurate simulation studies and better phase behaviour treatments.