The purpose of this study is to develop an automated selection method for cross‐sectional images with basal ganglia from brain CT scan. The proposed method includes binarization, morphological operation and labeling techniques. Area was calculated from the labeled region inside the cranial bones. The cross‐sectional images with basal ganglia were automatically selected using the calculated area. The method was applied to 22 cases each with 20 slices. As a result, the sensitivity and the specificity were 0.95 and 0.98 respectively. The proposed method would be a fundamental technique for the development of computer‐aided diagnosis（CAD）system.