Abstract: Conventionally the antenna selection algorithm provides high performance based on minimum bits error rate (BER) in spatial multiplexing system, but it needs the multiplication of a mass of channel matrices with high complexity in computation. To solve this problem, a low-complexity antenna selection algorithm was proposed on the basis of symbol error rate (SER). Studying the approximate equivalent relation between the signal-to-noise ratio (SNR) at receiving end and system channel matrices, an antenna selection algorithm is proposed utilizing the complex Householder-QR factorization. The proposed algorithm provides the high performance close to an optimal one, and has lower complexity in computation than conventional suboptimal algorithm. Simulation results demonstrated that the algorithm proposed enjoys favorable performance at lower cost in comparison with other algorithms.