Magnetic anomaly maps reflect the spatial distribution of magnetic sources, which may be located at different depths and have significantly different physical and geometrical properties, complicating the identification of the corresponding geologic structures. Filtering techniques are frequently used to balance anomalies from shallow and deep sources, and to enhance certain features of interest, such as the edges of the causative bodies. Most methods used for enhancing magnetic data are based on vertical or horizontal derivatives of the magnetic anomalies or combinations of them, and the edges or centers of the sources are identified by maxima, minima, or null values in the transformed data. Normalized derivatives methods are used to equalize signals from sources buried at different depths. We present an edge detector method for the enhancement of magnetic anomalies, which is based on the tilt angle of the total horizontal gradient. The notable features of this method are that it produces amplitude maxima over the source edges and that it equalizes signals from shallow and deep sources. The method is applied to synthetic and real data. The effectiveness of the method is evaluated by comparing it with other edge detection methods that have been previously reported in the literature and that make use of derivatives. The results show that our method is less sensitive to variations in the depth of the sources and that it indicates the position of the edges of causative bodies in a more accurate fashion, when compared with previous methods, even for anomalies due to multiple interfering sources. These results demonstrate that the proposed method is a useful tool for the qualitative interpretation of magnetic data.