Based on the background of complex coupled multiple tasks executed by heterogeneous multiple unmanned aerial vehicles(UAVs), a deadlock-free sequentially extended consensus based bundle algorithm is presented to solve the distributed task allocation problem. Firstly, a timing multi-task allocation model is established in consideration of the constraints of task load resources, task sequence and threat regions. Secondly, the task bundle construction and conflict resolution rules of the consensus based bundle algorithm are extended, and then a deadlock detection and correction method of the task scheme based on digraph depth-first search is designed to achieve conflict-free and deadlock-free task allocation. Furthermore, the timing constraints between related tasks are transformed into soft time window constraints, and the sequential hierarchical strategy is utilized to solve the problem. Finally, the Dubins curve path is employed to integrate the route planning into the task allocation in order to improve the reliability of the task allocation result. Simulation results show that the proposed algorithm can tackle the distributed coupled multi-task allocaiton problem of heterogeneous multiple UAVs quickly and efficiently, and possesses satisfactory optimality and timeliness..