In probabilistic risk assessment (PRA), thermal-hydraulic (TH) support studies are used to evaluate the consequences of accident sequences. These studies consist of analyst-selected simulations of plant evolution under accident conditions. The TH model describing the accident sequence is implemented in scripts that may be written using a different programming language depending on the TH software used. In order to guarantee the quality and the acceptability of the results the TH model, as implemented in the scripts, should be verified as well. The development of dynamic PRA (DPRA) methodologies introduces from this perspective a new challenge. In DPRA, each scenario is simulated using a verified and validated TH computer code. However, thousands of different simulations are run within a single analysis and using a single TH model implemented within a script that accounts for all the possible scenarios. Under these conditions, the verification techniques normally used can be ineffective. The use of a statechart tool is proposed that allows to visualize the model evolution as the different events occur and, consequently, checking the correctness of the model. The framework also allows to generate in an automatic way the model script from the statechart graphical model using a Python script. To illustrate the framework, MAAP5-EDF is used as the TH computer code, while the statechart is modeled using YAKINDU StateChart Tools software.