Abstarct: Using geophysical methods and measurement of physical properties of subsurface rocks is a suitable solution for exploration of underground reserves (oil, gas, water, minerals …). Gravity and magnetic surveys (potential field methods) are routine geophysical methods for initial stages of exploration, because of simplicity and low cost. There are several methods for automatic interpretation of potential field data. This study uses local phase filters for detection of anomaly boundaries and results will compare with other filters such as analytic signal, total horizontal derivative, and normalized standard deviation. All those necessary functions and codes were written using MATLAB software then, these filters apply to synthetic models. The results show that the normalized standard deviation filter has been successful in edge detection for gravity and magnetic data associated with synthetic models. Finally these filters were applied to gravity and magnetic data of the Saveh sedimentary basin as a case study. Here, similar to the synthetic data, normalized standard deviation filter shows better results than other filters and highlights structures such as faults, fold patterns and salt dome more clearly.