The use of artificial neural network (ANN) is becoming common in the analysis of hydrology and water resource problems. The method of ANN was used in this work to model flood occurrence in Coastal region of Cross River State, Nigeria. Meteorological data were collected from the Nigeria Meteorological Agency (NiMet) and Nigeria Hydrological Survey Agency (NHSA) for over 50 years. Artificial neural networks are known to have capacity for pattern recognition and have been proven to be reliable extrapolative tools for modelling of flood. The Levenberg-Marquardt backpropagation training algorithm was used with Nonlinear Autoregressive with external inputs (NARX) time series tool. Predictions from the neural network model were checked and validated using test of correlation coefficient (R), coefficient of determination (R²) and the mean square error (MSE) between the observations and predictions. The flood extrapolation from the neural network revealed an increase in water level of Calabar River in the months with extreme rainfall. This consequently in linear response with the catchment characteristics facilitated flooding.