Risk assessment of a thermal power station production unit and prediction of its working life is based on failure rates determined on the basis of available data and consequences due to malfunctioning of selected energetic devices. Probabilistic assessment is focused on selected critical devices including cooling tower, industrial chimney, boiler pressure system, steam piping, steam turbine, generator and transformer which failures lead to production loss. The risk assessment of a power station production unit is performed using the Bayesian networks. The results of risk analyses indicates that the devices with significant influence on economic risk of the selected production unit include the boiler pressure system, steam turbine and generator. It is shown that the societal risks are primarily affected by the steam turbine and steam piping.