The Wang-Landau Monte Carlo algorithm is implemented within an effective Hamiltonian approach and applied to BaTiO 3 bulk. The density of states obtained by this approach allows a highly accurate and straightforward calculation of various thermodynamic properties, including phase transition temperatures, as well as polarization, dielectric susceptibility, specific heat, and electrocaloric coefficient at any temperature. This approach yields rather smooth data even near phase transitions and provides direct access to entropy and free energy, which allow us to compute properties that are typically unaccessible by atomistic simulations. Examples of such latter properties are the nature (i.e., first order versus second order) of the phase transitions for different supercell sizes and the thermodynamic limit of the Curie temperature and latent heat.