In turbo-multiple-input multiple-output (Turbo-MIMO) systems, the soft-output MIMO detector can provide the priori information to the turbo decoder. Unfortunately, if Rayleigh fading channels are applied, the induced unreliable priori information would cause the system performance degradation. In this paper, we proposed an iterative method to acquire the high reliability priori information from MIMO soft-detector in Turbo-MIMO systems. Similar to the conventional updating rules in the turbo decoding algorithm, we utilize the extrinsic information from the turbo decoder to update the log-likelihood ratios (LLRs) based on log-MAP algorithm in the list sphere decoding (LSD) algorithm. To reduce the overall computational complexity, different iteration profiles are also discussed. Simulation results show that the proposed Turbo-MIMO system can significantly improve the system performance compared to that of the conventional Turbo-MIMO system.