Abstract: Multiple-input multiple-output (MIMO) and orthogonal frequency division multiplexing (OFDM) are the core technology of the new generation of communication system, the combination system called MIMO-OFDM can achieve higher capacity and overcome the frequency selective fading. The rank of channel matrix will be deficient as a result of antenna correlation or special scattering structure in wireless MIMO channels, in which the traditional MIMO detection algorithm cannot be applied directly. In order to overcome the problem, distributed receiving system was put forward, which can achieved full rank matrix required according to the extension of the channel matrix. The classical detection algorithms include QR decomposition algorithms and sphere decoding algorithms. In order to reach the detection results and reduce the complexity of the maximum likelihood, QRD-M detection algorithm was used to detect the sending signal of the distributed system above. To measure the result of MIMO detection, this paper gave the Cramer-Rao bound of the distributed system. Through the theoretical analysis, the CRB derivation and model simulation, this algorithm was proved to have the obvious advantages in against the channel rank loss compared with the existing rank-deficient MIMO detection algorithms, the algorithm perform well and the RMSE is more close to the CRB.