This study focuses on developing an automatic speech recognition system to classify Arabic sibilant consonants into alveolar and post-alveolar groups based on energy distribution as an acoustic cue. The method achieves high accuracy rates of 100% for overall classification, 96% for post-alveolar consonants, and over 94% for alveolar consonants, outperforming existing classification algorithms. The research contributes to the understanding and processing of Arabic speech sounds, utilizing acoustic analysis and machine learning techniques.