Two new classes of Boltzmann machines will be introduced in this paper. They are called Probabilistic Boltzmann Machine (PBM) and Automata Boltzmann Machine (ABM) by the author. The ABM is further divided into union ABM, intersection ABM, and mixed ABM. The motivation of this development is to enhance the capability of the Boltzmann machine family. These two new classes of machines are derived by combining the automata neural network developed by the author, the stochastic neural network developed by the author, and the Boltzmann machine.