The safety of Nuclear Power Plants (NPPs) is an important issue that is of concern to all including; regulators, operators, and the general public. Assuring the safety of a NPP is a primary objective by all stakeholders. In the wake of accidents in NPPs recorded in the past such as Chernobyl, TMI, and recently in Fukushima, the need to review existing safety system design and operation as well as performing safety verification of these systems as a means of preventing such accidents in the future is necessary. In this study, we present a framework for achieving safety verification of a NPP with emphasis on using cosimulation with reduced error for real time fault propagation analysis based on Fault Semantic Network in which a multiphysics model is mapped unto fault/risk models, and safety/protection systems of NPPs in order to achieve safety verification based on highest risks and previous accidents. A statistical method is used to reduce errors between simulation results and real time data which is illustrated with a case study from literature. (c) 2015 American Institute of Chemical Engineers