This paper describes a distributed approach for data fusion and information sharing based on evidence theory and the transferable belief model. Evidence theory aggregates data generated from different sources in order to better assess an ongoing situation and to aid in the response and decision making processes. In the domain of critical infrastructure protection, researchers are forced to develop distributed approaches for modeling and control with a minimal exchange of data due to the existence of multiple stakeholders and interconnections between infrastructure components. Evidence theory permits the modeling of uncertainty in data fusion, but it is typically applied in a centralized manner. This paper proposes a decentralized extension of the transferable belief model that facilitates the application of evidence theory to data fusion in critical infrastructure applications. A case study is provided to demonstrate the convergence of results similar to the centralized approach, and to show the utility of fusing data in a distributed manner for interdependent critical infrastructure systems.