Abstract: The standard nearest-neighbor data association algorithm may generate miss-tracking and lose target in a clutter environment. To handle this problem, this paper proposes a reverse prediction weighted neighbor data association algorithm for considering all candidate echoes. After calculating the candidate echoes'reverse prediction residual norm, the normalized weight for each candidate echo is obtained. The equivalent echo that is weighted sum of candidate echoes is used to update the target state. The algorithm effectively reduces the error association by using equivalent echo. The simulation results show that the algorithm can keep less amount of calculation and is effective to avoid miss-tracking and lose target.