In sonar systems, the reverberation and clutter is a primary cause of false alarm. The multiple return association and localization (MRAL) algorithm is a data-fusion algorithm that can efficiently reduce the false alarm rate. However, MRAL does not show which one of outputs is from a target. In this paper, we propose an algorithm that find which one of outputs is from a target by finding heading angle and speed variances of received signals using Doppler shift frequency.