Aiming at the global task planning problem of large-scale heterogeneous UAV (unmanned aerial vehicle) swarm, a task planning method based on balanced clustering market auction mechanism was proposed. Scene of completing tasks by collaborative UAVs was analyzed, and a task planning model with high generality was established by combining the advantages of task clustering and UAV coalition. Considering the demand for load balance of UAVs, a new balanced clustering market auction algorithm which comprehensively considers the travel consumption and task consumption was established by integrating and improving the K-means algorithm and market auction mechanism. The balance parameter was introduced into the auction process. By solving the traveling salesman problem to modify the balance parameter, the total cost was continuously reduced while ensuring the load balance. The simulation results show that the task planning method using balanced clustering market auction mechanism can complete the complex task planning of heterogeneous UAV swarm in a short time, ensures the load balance of UAV coalitions, and has good performance in total cost and total time, exhibiting certain practical application value.