Liquid-liquid extraction (LLE) is an important technique to separate aromatics from aliphatics since these compounds have very similar boiling points and cannot be separated by distillation. Ionic liquids (ILs) are considered as potential extractants to extract aromatics from aliphatics. In this paper, molecular dynamics (MD) simulations were used to predict the extraction property (i.e., capacity and selectivity) of ILs for the LLE of aromatics from aliphatics. The extraction properties of seven different ILs including [C2mim][Tf2N], [C2mim][TFO], [C2mim][SCN], [C2mim][DCA], [C2mim][TCM], [C4mim][Tf2N], and [C8mim][Tf2N] were investigated. Results show that ILs with shorter alkyl chain cations and [Tf2N]- anion exhibit better extraction efficiency than other ILs, which is in agreement with previously reported experimental data on the extraction of toluene from aliphatics and further validated the reliability of the proposed model. The binding energies between ILs and organic molecules were calculated by the density functional theory, which help explain the different extraction behaviors of different ILs. The symmetry-adapted perturbation theory analysis was performed to further understand the interaction mechanisms between ILs and organics. Our study shows that the [Tf2N]- anion also has the best extraction capability for heavier aromatics (o-xylene, m-xylene, and p-xylene) from common aliphatics (heptane and octane). The MD modeling approach can be a low-cost in silico tool for the high-throughput fast screening of ILs for the LLE of aromatics from aliphatics.