In order to solve the problem of two-dimensional quantum circuit qubit nearest neighbor constraint in some quantum techniques, a priority-based nearest neighbor interaction cost measurement model is proposed. The optimal layout of the qubit in two-dimensional architecture is obtained based on the harmony search (HS) algorithm and then the insertion of the SWAP gate is performed by the given local sorting method. Finally the quantum circuit realizes the nearest neighbor interaction under the two-dimensional architecture. The proposed algorithm is verified by experiments and compared with the latest related results. The experimental results show that the SWAP gates in proposed method is reduced by 14.42% on average compared with that in the two-dimensional grid architecture reported in the literatures.