Thai Classical Music has its own standards for pitch frequency and frequency range that differ from Western music. Therefore, applying Western note identification algorithms to Thai Classical Music results in inaccurate note identification, especially when used for creating musical signatures for retrieval or similarity comparison. Furthermore, the pitch frequencies of Thai Classical Music vary across different standards. This research proposes a new algorithm for identifying notes in Thai Classical Music based on a regression method. Using a dataset compiled from textbooks and musical pieces from various sources, the researchers derived an equation to identify the pitch frequencies of Thai musical notes. This equation was then used to generate a new frequency table, referencing the sequence of notes on a piano. The signature creation process for retrieval begins by segmenting the audio data into smaller parts, then applying the Fast Fourier Transform (FFT) to convert each segment into frequency data. The peak frequency is extracted from the one-sideband frequency data is analyzed to identify peak frequencies, which are then matched with a proposed frequency table to determine the corresponding musical notes. These notes are then concatenated into a character string, forming the signature of the piece. Experimental results from developing a music retrieval system show that the proposed method outperforms the traditional Western-based method, achieving an accuracy of 72.53 percent.