This paper proposes methodologies for emergency flood forecasting by using neural network models.The success of methodologies will ensure the safety of human lives and the reduction of damage to the environment.Our study investigated the prediction potential on the actual data from Chao Phraya River at Bangkok for the year 2002.We have examined not one but several neural network models.The study indicated that Neural Network models can be used for forecasting flooding within the next 48 hours more accurately than the previous classical linear model.