The structural integrity of a reactor pressure vessel (RPV) is a crucial issue for an operating nuclear power plant, especially in the beltline region, which suffers the highest neutron irradiation. Owing to its capability of considering parameters based on statistical distributions and provision of objective risk-informed results, the probabilistic fracture mechanics (PFM) method is widely used in evaluating the structural integrity of RPVs. However, the flaw characteristics used for PFM analysis are mainly derived from particular vessel inspection information such as from the Pressure Vessel Research User Facility and Shoreham vessels, which may not be able to truly represent the vessel-specific condition of an analyzed RPV. In this work, the Bayesian inference, which combines prior flaw data with new inspection results as well as uncertainties, is used to develop posterior vessel-specific flaw distributions. Then, the updated flaw model is used for PFM analysis to investigate the effects on the fracture probability assessment of RPVs subjected to pressurized thermal shocks (PTSs). Considering the updated flaws based on the inspection data, the PFM analysis result could be more realistic to predict the fracture risks of RPVs during operation.