Abstract: In order to improve the search accuracy of twin sequences and reduce the computational cost, a new K nearest neighbor search method based on Chebyshev distance is proposed. A tree-based index TI was introduced, which pruned the search space by using the appropriate boundaries in its nodes. An optimization method was introduced to improve the memory of the index by batch loading, so as to prevent the performance from deteriorating and reduce its search cost. Based on Chebyshev distance, the twin sequence K nearest neighbor search algorithm was implemented on the index. The experimental results show that the proposed method can achieve efficient and high-precision search.