A methodology is described that can be used for the extrapolation of thermal-hydraulic phenomena measured in differently scaled integral test facilities to nuclear reactor plant conditions. The use of a system code in this context is confirmed to be of fundamental importance, provided that the code’s scaling capability has been demonstrated. The starting data base for the proposed study consists of the measured quantities and corresponding RELAP5/MOD2 code calculation results related to a boiling water reactor small-break loss-of-coolant accident (SBLOCA) counterpart test activity, a pressurized water reactor (PWR) natural-circulation type test activity, and a PWR SBLOCA counterpart test activity. The proof that this methodology can be used for evaluating uncertainties in predicting transient behavior in nuclear power plants is the main result of this study. Data have been obtained that give a value of the foreseeable error ranges in the provision of plant behavior in the three cases considered.