Abstarct: The concept of intervals in Persian music and its effects on music classification and configuration is one of the most important issues considered by musicians. It has also been taken into account in signal processing applications to achieve automatic analytic methods. Mode (scale) detection from music data directly relates to the concept of intervals. One of the most applicable features, used in mode detection systems, is pitch. Such a melodic feature, besides other melodic and rhythmic features, is used for classification and different musical similarity related analysis in the area of Music Information Retrieval (MIR). In this thesis, a novel algorithm for recognizing different modes of Persian music is proposed. For mode detection, a 24-pitch profile feature, which contains essential information about the mode of the piece of music, has been extracted from the musical audio. Moreover, for feature extraction, a dimension reduction step was proposed, which obviously reduced the calculations and improved the speed of the algorithm. Finally the experiments on the collected databaxse showed promising results for Persian music mode detection.