This work presents a molecular thermodynamic approach in which microemulsion properties are obtained by minimizing the total Gibbs energy of the system. Notably, this optimization process is carried out mainly based on the surfactant molecular structure and the solution conditions. Even though molecular thermodynamic modeling has been widely applied to oil-in-water (O/W) microemulsions, just a few studies are reported on water-in-oil (W/O) microemulsions mainly focused on ionic surfactants with linear alkyl tails. Here, we investigate the self-aggregation of different non-ionic surfactants with more complex structures acting as hydrophobic tails. In this context, we model the phase behavior of alkylphenol ethoxylate surfactant systems incorporating liquid-vapor and liquid-liquid data of alkylbenzenes from the literature. As a result, we can optimize the model parameters and describe the equilibrium micelle structure. The particle swarm optimization method was applied to minimize the Gibbs free energy for both O/W and W/O microemulsions. In order to evaluate the predictive performance of the model, we compared the experimental data with the computed critical micelle concentration and micelle size.