The task assignment problem for multiple UAVs is highly nonlinear and confrontational, with significant practical applications in multi-agent systems. In this paper, we study the task assignment problem involving isomorphic multiple UAVs and multiple stationary targets. Building on the Multiple Traveling Salesperson Problem (MTSP) model, we enhance the traditional genetic algorithm. Initially, we adopt a bidirectional coding method using distribution vectors and breakpoint vectors to simplify the routine coding process. Subsequently, a group selection strategy is introduced to select individuals, and a class 3-opt local search strategy is employed to find the optimal solution vector, which improves the convergence speed and global search capability of the algorithm. We propose an improved genetic algorithm based on the Group Local Search strategy (GLS-IGA). Experimental results demonstrate that the proposed GLS-IGA performs better in the multi-UAV reconnaissance task assignment problem.