In this paper, we attempted to determine the location of fault zone using the Genetic Celluar Neular Network Method (G-CNN) in the Thrace and the Marmara Sea regions. G-CNN is a method used to detect specific feature of the 2-D image in the image-processing technic. Genetic algorithm has been used for as learning algorithm in the G-CNN. The G-CNN method has been used for determining the fault zone, as detect regional and residual effect of the gravity anomaly map. Regional anomaly map has been modelled accordingly and compared to the avaliable seismic data. The fault zones in these regions, have been determined by way of companion of the fault model with geological data the outcome of which ultimately gives complete accord.