Identifying the boundaries of source bodies is one of the most important aims in interpreting gravity and magnetic data. For this purpose, many amplitude and phase-based methods have been utilized. Amplitude-based filters cannot balance the anomalies caused by deep and shallow sources. Although phase-based filters do not have such a problem, the edges detected by using these filters may be incompatible with actual edges. To overcome these problems, many maximum detection algorithms have been developed, e.g., the curvature analysis algorithm, the Blakely and Simpson's algorithm, and its modified version. In this study, the edge mapping performances of these algorithms are discussed and a new approach is introduced. The validity of the introduced algorithm is illustrated with two synthetic gravity data caused by vertical-prism sources. The results show that the proposed technique detects the source edges more clearly. In addition, the proposed technique has been applied to the gravity data of the Tuan Giao area in Vietnam and the magnetic data of the Narmada-Son Lineament in India, and it has been observed that the obtained results are consistent with the known geological structures in these study areas.