Problems associated with flood risk definitely represent a highly topical issue in Vietnam. The case of the lower Gianh River in the central area of Vietnam, with a watershed area of 353 km2, is particularly interesting. In this area, periodically subject to flood risk, the scientific question is strongly linked to risk management. In addition, flood risk is the consequence of the hydrological hazard of an event and the damages related to this event. For this reason, our approach is based on hydrodynamic modelling using Mike Flood to simulate the runoff during a flood event. Unfortunately the data in the studied area are quite limited. Our computation of the flood risk is based on a three-step modelling process, using rainfall data coming from 8 stations, cross sections, the topographic map and the land-use map. The first step consists of creating a 1-D model using Mike 11, in order to simulate the runoff in the minor river bed. In the second step, we use Mike 21 to create a 2-D model to simulate the runoff in the flood plain. The last step allows us to couple the two models in order to precisely describe the variables for the hazard analysis in the flood plain (the water level, the speed, the extent of the flooding). Moreover the model is calibrated and verified using observational data of the water level at hydrologic stations and field control data (on the one hand flood height measurements, on the other hand interviews with the community and with the local councillors). We then generate GIS maps in order to improve flood hazard management, which allows us to create flood hazard maps by coupling the flood plain map and the runoff speed map. Our results show that: the flood peak, caused by typhoon Nari, reached more than 6 m on October 16th 2013 at 4 p.m. (its area was extended by 149 km²). End that the typhoon constitutes an extreme flood hazard for 11.39%, very high for 10.60%, high for 30.79%, medium for 31.91% and a light flood hazard for 15.32% of the flood area.