Vesicles and micelle electrokinetic chromatography (EKC) were used to study solute partitioning from the aqueous phase into vesicles and micelles. The retention factors (k) of neutral solutes are related to their partition coefficients (K) and the phase ratio (f), k = K*f. Dihexadecyl phosphate (DHP) vesicles used in this study undergo gel to liquid-crystalline phase at the critical temperature. Investigation of thermodynamics of solute partitioning into bilayers would allow separating the enthalpic and entropic contributions to the free energy of transfer. Linear Solvation Energy Relationship (LSER) modeling was used to elucidate the contributions of hydrophobic, hydrogen bond interactions, dipolar, and polarizability to the free energy of transfer of solutes from the aqueous phase into the vesicle phase at above and below Tc. Hydrophobic interaction keeps an important role in solute partitioning into bilayers. Hydrogen bonding, electrostatic, and dipolar are also relative important. Polarity (p*) was directly measured at different locations in the DHP vesicles using different solvatochromic probes over a wide range of temperature. The size of liposomes and vesicles were also examined at different temperatures. In peptide mapping, models for the prediction of electrophoretic mobility and micelle-water partition coefficient of peptides were developed, and thus retention behavior of peptides, which should facilitate method development in MEKC. It allows rapid optimization of separation conditions that can lead to enhance resolution of complex mixtures. Initially, the partition coefficients of a training set of peptides were determined by MEKC. A quantitative structure-partition relationship (QSPR) was established based on the data set that relates partition coefficient (Kmc) to structural descriptors of the peptides. Also, a quantitative structure-migration relationship (QSMR) was developed based on the charge, residue mass, and length of peptide. An advantage of this model is that peptide descriptors can be calculated from the amino acid composition of the peptides. Different statistical methods were then used to determine the most relevant descriptors in the models and to test the accuracy and error for each model. The QSPR and QSMR models were then used to predict Kmc and peptide mobility for several membrane peptides and tryptic digest of horse cytochrome C.