To address the challenges of high positioning accuracy requirements in unmanned aerial vehicle swarm localization， a cluster-based cooperative localization algorithm for unmanned aerial vehicles based on difference-of-convex （DC） programming is proposed. A cooperative localization mathematical model based on positional information is established. Then， following the structure of convex optimization algorithms， the non-convex constraints are transformed into a difference-of-convex objective function to enhance the accuracy of the position solution. To optimize the iterative process， the multi-dimensional scaling （MDS） method is used to provide an initial position estimate， and this initialization step is incorporated into each round of intra-cluster localization. The algorithm is further analyzed under the condition of ranging errors. The objective function is reformulated using maximum likelihood estimation to reduce positioning errors and improve location estimation accuracy. Subsequently， a cluster fusion method is introduced， which utilizes common nodes and the Procrustes analysis algorithm to achieve global localization. Simulation experiments are conducted to measure localization errors， and comparisons with commonly used localization algorithms are made to validate the effectiveness of the proposed method. The results demonstrate that the proposed algorithm features high localization accuracy， and is well-suited for multi-unmanned aerial vehicle networks， effectively enhancing the localization performance of UAV swarms.