This tutorial provides an overview of Bayesian belief networks. The subject is introduced through a discussion on probabilistic models that covers probability language, dependency models, graphical representations of models, and belief networks as a particular representation of probabilistic models. The general class of causal belief networks is presented, and the concept of d-separation and its relationship with independence in probabilistic models is introduced. This leads to a description of Bayesian belief networks as a specific class of causal belief networks, with detailed discussion on belief propagation and practical network design. The target recognition problem is presented as an example of the application of Bayesian belief networks to a real problem, and the tutorial concludes with a brief summary of Bayesian belief networks. A Tutorial on Bayesian Belief Networks