Abstract: In this paper, an improved codeword searching algorithm is proposed on the basis of nearest-neighbor search algorithm. The new algorithm considers the sums and variances of image vectors. A vector is separated into two subvectors: the first half of the coordinates and the second half of the coordinates. Calculate the sums and variances of the vector and its two subvectors. Apply the result to a set of inequalities to eliminate the impossible codeword candidates. The simulation results show that the proposed algorithm is faster than the improved EENNS algorithm, and it also has the advantage in decreasing the computing complexity.