Recently, as the progress of information- communication technologies, lots of valuable data related to water environment are automatically observed and some of them are opened to public in real time through media such as internet etc. It is expected that these data will contribute to the solving problems occurred at real water areas.
In this study, regarding the facts that sufficient data are recently available to realize the data- driven modelling, genetic programming was newly developed and its applicability to the prediction of sea level departure was discussed with comparing to the results obtained from artificial neural network model. Although the training of genetic programming was executed for limited data, construction of tree was successfully established by regarded genetic procedures. Obtained tree considerably simulates the temporal changes of sea level anomaly observed at Kuchinotsu port. While, for obtaining the tree which outputs results with smaller errors, computational conditions, such as observed data fed to model and parameters on genetic procedures, have to be chosen with care.