This dissertation presents the development of two different types of polarization methods for molecular simulation methods, including Monte Carlo and molecular dynamics (MD) simulations. The first model, which is a polarizable intermolecular potential function (PIPF) method, is based on the point dipole method, where polarization energy is obtained from induced dipole moments and is added as correction to a force field. Hydrogen sulfide (H2S) molecule is studied and parameterized for the PIPF method, and this study displays that the PIPF method reproduces experimental gas-phase dipole moment, molecular polarizability, liquid density, and heat of vaporization very well with a relative error of less than 1.0%. Due to the over-polarization of the model, however, some liquid properties and liquid structure failed to reproduce experimental values, which indicates further improvement is necessary for the PIPF method. The second one is an explicit polarization (X-Pol) method, which is a self-consistent fragment-based electronic structure theory in which molecular orbitals are block-localized within fragments of a cluster, macromolecule, or condensed-phase system. The Lennard-Jone potential function is incorporated into the X-Pol potential in order to express short-range exchange repulsion and long-range dispersion interactions. The X-Pol potential is first developed at the B3LYP hybrid density functional with the 6-31G(d) basis set, and the Lennard-Jones parameters have been optimized on a dataset consisting of 105 hydrogen-bonded bimolecular complexes. It is shown that the X-Pol potential can be optimized to provide a good description of hydrogen bonding interactions; the root mean square deviation (RMSD) of the computed binding energies from CCSD(T)/aug-ccpVDZ results is 0.8 kcal/mol, and that of the calculated hydrogen bond distances is about 0.1 Å from B3LYP/aug-cc-pVDZ optimizations. In addition, the explicit polarization with three-point charge potential (XP3P) model is introduced using the polarized molecular orbital model for water (PMOw). The XP3P model is shown to be suitable for modeling both gas-phase clusters and liquid water, which is demonstrated from simulations of gas-phase water and protonated water clusters, and pure liquid consisting of 267 water molecules in a periodic system. This model is anticipated to be useful for simulating biological system in the condensed phase.