Fibroa Adenoma Mammae (FAM) or called a benign tumor is the most common tumor found in the breast, often this disease is considered as breast cancer by some laymen, but this disease is different from cancer because Fibroadenoma (FAM) can grow in all parts of the breast, In identifying FAM, doctors or radiologists usually have to analyze carefully the images of Magnetic Resonance stored in the format of Digital Imaging Communication In Medicine (DICOM). This process is certainly quite time consuming. Thus the author feels the need to create a digital image processing application to help doctors or radiologists identify FAM in the breast by utilizing the segmentation process in medical images of USG results using the Sobel method. This method performs the original image segmentation process by detecting the edges of the image then the segmentation results are converted into binary images so the system can determine the FAM area. Medical image segmentation using the Sobel method is good for determining the edges of FAM objects because the edges can be clearly seen but for some image images with less resolution as previously tested, edge detection will be difficult to determine the edges of smooth objects and only form lines rough edges.