Quantification of the reliability of systems is a important task when qualifying the new technology, since a lack of adequate reliability performance will violate the intended gain of the innovation. Several methods for model-based assessment have been proposed in literatures. However, they are often be criticized for lacking of ability to fully address the complicated system performance. It is especially a case for the 'low probability but high consequence' subsea system where the cost element (including investment and consequence impact) are sensitive to the accurance of reliability quantification. Bayesian formalism, as a probabilistic approach, is experiencing a growing success in the reliability assessment due to its flexibility. This paper reviews the advantages offered by Bayesian formalism, and explore the possible use of Bayesian Network in quantifying reliability with respect to new or future subsea systems, where few data available to support the quantification and the performance of system is often too complex to comprehend in the commonly used model. The applicability of Bayesian Network is demonstrated by the example of high integrity pressure protection system.