The expression and recognition of geographical environment, the recovery and reappearance of ancient geographical environment have been the focus in some research fields such as geography, cartography, cognitive science and artificial intelligence. As well, virtual geographical environment and the expression of multidimensional spatio-temporal information have become important research directions with the development of GIS. The reconstruction of sea surface change has long been a focus of geography, and in the past years the prediction of sea surface change has also received considerable interest. The recent development of GIS provides the opportunity for the dynamical simulations of sea surface change as well as for its visualizations. It is also an important research direction in the development situation to build the model with functions of forecast, influence and countermeasure, based on the research of the sea level historical change. This model can provide an effective method for sea level change research. In this paper we present an adaptive 3D Old Sea Level-Land Surface Change Simulation Model based on GIS for the above purpose. This model can be used to simulate the old sea level-land surface evolution over coastal regions. With historical records and expertise knowledge, this model automatically verifies the simulation results and adjusts the parameters which controls the simulation; and then iteratively repeats the simulations until the results become close to the recorded history of old sea level-land surface evolution. Compared with other models, this model has the following features: relatively accurately controlling calculation, easily repeating simulations, automatically verifying the calculation results and automatically/semi-automatically adjusting the parameters. We used this model to simulate the old sea level-land surface change of the Yangtze River delta in the past ten thousand years, and it demonstrated pretty good performance.