Aeromagnetic lineaments are used to identify and recognize geological structures. There are Different methods to identify magnetic lineaments which are generally based on horizontal and vertical gradients. These methods are sensitive to noise. Nowadays, machine learning methods for solving geophysical problems have been developed and Its application is increasing. In this research, the Convolutional Neural Network (CNN) method, which is a subset of machine learning methods, is used to determine the aeromagnetic lineaments of eastern Iran. This method was developed by Smith and Naprstek (2022). From the information obtained by lineaments and geological information, The faults and tectonic boundaries of eastern Iran have been interpreted. The extracted magnetic lineaments have a good match with the linear structures of the region, including Doruneh, Nehbandan, Naybandan, Uzbak-Kuh, Biabanak, Zarand faults. In addition, according to the faults, the boundaries of Lut block, Tabas block, Yazd block, Kashmar-Kerman tectonic zone have been determined.