A new channel estimation approach for orthogonally coded multiple input multiple output (MIMO) systemis proposed. The main idea of the technique proposed in this study involves the use of specific properties ofthe orthogonal space-time block codes (OSTBCs) in the estimation based on the second order statistic (SOS)to estimate the channel matrix after the application of singular value decomposition (SVD). This modifiedsolution directly brings simple parameters in comparison with conventional estimation from the channel matrix.In this investigation, the authors made an attempt to reduce the effect of ambiguity which could be seen inmost of the traditional blind algorithms. As the training-based least square (LS) was still used for a part of thechannel matrix, the scheme in this study could be described as the semi-blind algorithm which has stronglyattracted more researchers in the wireless field. A detailed analysis carried out in this current research workhas validated the performance and computational advantages of the proposed method. Meanwhile, results fromthe simulation have confirmed the modified scheme and illustrated that the new semi-blind channel estimationis capable of improving the quality of the overall system.