Although multiple-input-multiple-output (MIMO) detection has received much research attention in the past years, to the authors knowledge, few detection methods demonstrate optimal/near-optimal performance with low complexity. This paper proposes to incorporate automatic retransmission request (ARQ) with sub-optimal MIMO detectors so as to achieve both favorable performance and low complexity. In the study, retransmission delay induced by ARQ is exploited as a source of improving the detection performance of Iow complexity algorithms. In particular, the detection performance of sub-optimal algorithms improved by introducing ARQ is analyzed theoretically. A sufficient condition for such scheme to achieve full-diversity performance is also derived which relates detection performance with number of transmission times. Moreover,throughput cost by retransmission is deduced as well as its lower bound. The zero-forcing (ZF) equalizer cooperating with ARQ,as a case study, is shown to have evident performance improvement through theoretical analysis. And numerical results are presented to verify the effectiveness of the proposed scheme which boosts the performance of sub-optimal detector and possesses lower implementation complexity for practical reality simultaneously.