Abstarct: Detection of traditional Iranian music classics for people without specialization or recruiting students is very difficult and often impossible. The conceptual recovery of music information is baxsed on the intelligent and automatic analysis of music, with the aim of easy access to music or information, which recognizes the music scales of one of the uses of conceptual information retrieval of Iranian music. One of the features of the music signal, which is very important for processing in the field of music, is the Mel frequency cepstrum coefficient. In this thesis, by studying the existing methods in the field of Western and non-Western information retrieval and considering the fundamental structural differences between traditional Iranian music and other music, especially western music, a new method It is proposed to observe the special considerations of traditional Persian music and recognize the type of position of each piece of music with a high percentage. two-laxyer perceptron neural network has been used for detecting. By providing two traditional Iranian music databaxses of experienced and experienced teachers with different tools, by extracting the characteristics and applying the perceptron neural network to the databaxses, the average percentage accuracy of detection was about 98%, which compared with similar methods, it has a lot precision.