Hydrogen Stress Cracking (HSC) of steels and CRA has been studied for a long time and it is recognized that the results have significant variability depending on many factors, including test procedures, material microstructures, and environmental variables. Although the standards used for acceptance of a material for sour service are deterministic, the variabilities in the test data as well as field conditions merit probabilistic treatment. This paper provides a Bayesian network model of HSC that connects the variabilities in material, environment, and loading conditions through mechanistic factors related to hydrogen transport, deformation mode, and strain rate. The model tries to encapsulate the state of knowledge in this area and tests the outcomes in terms of probability of HSC. The sensitivity of the model to input variables is assessed.