Abstarct: Since fault has great significances for the trapping and migration of hydrocarbon resources, fault detection is one of the most important components of seismic interpretation in petroleum exploration. Traditionally, Fault interpretation is done by inspecting seismic sections and time slices by interpreter. This method is subjective and also time consuming. Afterwards, because of noise effect and also weak reflections, fault pattern visualization and detection on the seismic data is a sensitive task and challenging problem. Therefore, using seismic attributes for this problem has been widely developed in recent decades. Gorgan plain in the north of Iran is an unknown region about hydrocarbon exploration and a few researches have been done in this area. In this thesis a research has been done about delineating the fault pattern in the seismic data acquired on this region. The research method is baxsed on using some appropriate seismic attributes in fault detection. The most of appropriate attributes for fault detection are baxsed on discontinuity and dissimilarity along reflectors on seismic data. Several attributes were tested and finally four attributes were selected for the purpose of research. These attributes are variance, amplitude contrast, chaos and similarity. The attributes analysis has been done on five 2D seismic lines. The results have shown that the variance attribute has a superior efficiency in fault pattern visualization and the weakest results have been provided by chaos attribute. Afterwards, in this research, a structural model using existing information about subsurface formations and also by picking reflectors on seismic data has been provided which the detected fault patterns by seismic attributes have been shown in this model.