Fluorous biphase system (FBS) exhibits significant prospect in synthesis and catalysis fields. With the development of FBS, the design and prediction on the fluorophilicity of compounds show major importance. QSPR models are established by using threedimensional structural descriptors derived from VolSurf program and two-dimensional topological descriptors derived from Tsar program. Good results are received and the training set model has satisfactory predictive capability for the test set. MLR method is applicative for Tsar descriptors and the equation shows that more number of F atoms in the compound will result in higher fluorophilicity. PLS analysis is suitable for VolSurf descriptors. Analysis on VolSurf descriptors demonstrates that large hydrophilic regions at medium energy level are beneficial to the fluorophilicity. The capacity factor is negative correlation with the fluorophilicity, so do the best local interaction energy minimum and its distance. High hydrophobic descriptor values and large hydrogen bond acceptor regions at medium energy levels are favorable elements. High anionic regions, low POL value and high MW are advantageous to the fluorophilicity. 3D-VolSurf descriptors (R 2 ¼ 0.976, Q 2 ¼ 0.966) are more satisfying and suitable to build the QSPR model for the fluorophilicity than other descriptors. In addition, chirality has different extent influence on the calculated result for the different chiral molecules or with the different modeling methods. Using single configuration to compute for chiral molecules will be hopeful to bring about more accurate and credible result.
