In this letter, we investigate the UAV swarm communication in the presence of jammers, where cluster-based network architecture is adopted to enable efficient backhaul links. We aim to maximize the sum rate of the cluster head (CH) UAV by optimizing UAVs' trajectories and clustering jointly subject to UAV's various flying abilities. However, the formulated problem is non-convex and thus intractable to solve. Thus, based on block coordinate descent (BCD) and successive convex approximation (SCA) techniques, an iterative algorithm is proposed to solve this problem. Numerical results show that the proposed algorithm outperforms the benchmark algorithm for jointly designing the trajectory and clustering. Moreover, it also shows that the fixed-wing UAV is more favorable in anti-jamming UAV swarm communications, giving a certain practical significance.