Verification and validation analyses of the RETRAN-03 code have been conducted by simulating the startup tests of a BWR-5 plant with the measured data from the 1100-MW(electric) boiling water reactor (BWR) of Tokyo Electric Power Company's nuclear power station. The comparison of results calculated by RETRAN-03 to BWR-5 measured data shows good agreement in tests changing reactor pressure and water-level setpoints and in primary-loop-recirculation (PLR) pump trip, main steam isolation valve (MSIV) closure, and generator load rejection tests. The calculated water level behavior could be well tracked with BWR-5 data. The effect of initial mixed-water level in a separator on the prediction of reactor water level was examined. Sensitivity analyses for the four- and the five-equation models were performed for a one-PLR-pump trip test, an all-MSIV closure test, and a generator load rejection test with 100% bypass. Such mild transients as startup tests show no distinct difference between the four- and the five-equation models. The various kinds of simulation with the start-up phase tests in changing pressure, water level, and power by transient could verify the RETRAN-03 model and validate the code.