The fuzzy probability rough set under the two universes information system is an important extension model of rough set theory. However， there is currently no research on the three-way decision aspects of this model. To address this issue， a fuzzy probability three-way decision model in the two universes information system is proposed in this paper. Firstly， the concepts of decision actions and decision costs are defined based on the two universes fuzzy probability rough set； then， based on minimizing the cost of Bayesian decision-making， a three-way partition of decision regions under the two universes fuzzy probability relation was derived， which is the three-way decision model； finally， a fuzzy probability three-way decision classification algorithm for the two universes information systems was proposed. The experiments has demonstrated the effectiveness and superiority of the proposed three-way decision classification algorithms on the two universes information systems.