Abstract: Based on the analysis of K-means algorithm and cuckoo search (CS) algorithm, this paper proposes a clustering algorithm based on improved cuckoo search algorithm. The algorithm combines the strong local search ability provided by K-means algorithm with the strong global search ability provided by improved cuckoo search algorithm. Which not only improves the local search ability of the cuckoo search algorithm, but also accelerates the convergence rate. At the same time, because the cuckoo search algorithm has strong global search ability, it can effectively prevent the premature convergence phenomenon, so it can effectively achieve the global optimal result. Experiments show that the clustering algorithm has a better convergence effect.