Fractal is characterized by scale invariance which means a self-similarity statistically in a certain range of scales. Self-similarity is very common in geological phenomena, such as structural shape, deposit distribution, and element content of rock. In this paper, taking the Duobaoshan district as an example, the component reflecting iron anomalies is calculated by principal component analysis and the threshold of anomaly gray is determined by a fractal model. Then, the authors compare the iron anomalies, △T contours, Bouguer gravity anomaly contours and known deposits. The result shows that the extracted iron anomalies are consistent with the faults and known ore deposits.