Abstract: This study presents lake Kerkini water level simulation. Water level depends on a large number of parameters and procedures which are usually complex or non-linear. Water level was calculated, by using a model based on visual basic language. The model took account of all parameters that contribute to water level. Simulation was achieved when the model output approximated the available measured values. Afterwards, the same project was implemented by using artificial intelligence methods. These are, artificial neural networks and adaptive neuro fuzzy inference system. The basic advantage of this implementation is the fact that the output is obtained without having to use all the parameters that contribute to the final result. This means that they can be implemented for modeling systems where the procedures are not fully known or when there is a large parameter number affecting the result. Both models showed a great performance in simulating water level fluctuation and they are also suggested for prediction.