<p>MPM is an effective method for prioritizing exploration targets, and its popularity has increased with the advancement of machine learning techniques. Deep learning, a subfield of machine learning, has recently gained attention in MPM. In the present study, the convolutional neural network (CNN) was applied as a deep learning approach for porphyry copper prospectivity modeling in the Shahr-e-Babak study area. The performance of the CNN method was compared with that of the Random Forest (RF) algorithm, which is a widely used and well-established technique in MPM. In total, 37 Cu indices and ten exploration layers were utilized in the modeling process. The evidential layers were derived from geological, remote sensing, stream sediment, and airborne geophysical data. After generating the porphyry copper potential models, a confusion matrix was used to compare their performance. The classification accuracy of the CNN method was 88%, while that of the RF method was 92%. Furthermore, the sensitivity, overall accuracy, positive prediction rate, and negative prediction rate for the CNN were 93%, 89%, 93%, and 85%, respectively. For the RF method, these parameters were 93%, 90%, 93%, and 90%, respectively. A fractal analysis was employed to distinguish background values from high-potential zones in both models. Model performance was further evaluated using the receiver operating characteristic (ROC) curve, yielding area under the curve (AUC) values of 0.98 for the RF model and 0.97 for the CNN model. Overall, the results indicate that the CNN deep learning method performed favorably in identifying porphyry copper mineralization, demonstrating its potential as a reliable tool for mineral prospectivity mapping.</p>