Abstract: The sparsity of concatenated convolutional channel in two-way amplify-and-forward relay was analyzed. Through the appropriate design of the pilot, the synthetic concatenated convolutional channels could be decomposed into two independent concatenated convolutional channels for channel estimation. Theoretical analysis and simulation results show that the concatenated convolutional channel possesses sparsity which varies within a range. The proposed scheme completes the two-way exchange of information by finishing concatenated convolutional channel estimation in end nodes. Without the priori information of the channel sparsity, this scheme improves the utilization of spectrum resources and the performance of channel estimation.