This paper addresses the problem of optimizing multi-user communication scheduling, association, UAV trajectory, and transmit power control in a multi-UAV enabled wireless communication system. The goal is to maximize the minimum throughput among all ground users by jointly optimizing these variables. The problem is formulated as a mixed-integer non-convex optimization problem, which is challenging to solve. To tackle this, an efficient iterative algorithm is proposed, leveraging block coordinate descent and successive convex optimization techniques. The algorithm alternately optimizes user scheduling and association, UAV trajectory, and transmit power, with each iteration focusing on one block while keeping the others fixed. For the non-convex trajectory and power control problems, approximate convex optimization techniques are applied. Additionally, a low-complexity initialization scheme based on simple circular trajectories and circle packing is proposed to speed up convergence. Numerical results demonstrate significant throughput gains compared to static UAV or other benchmark schemes, highlighting the effectiveness of the proposed design.This paper addresses the problem of optimizing multi-user communication scheduling, association, UAV trajectory, and transmit power control in a multi-UAV enabled wireless communication system. The goal is to maximize the minimum throughput among all ground users by jointly optimizing these variables. The problem is formulated as a mixed-integer non-convex optimization problem, which is challenging to solve. To tackle this, an efficient iterative algorithm is proposed, leveraging block coordinate descent and successive convex optimization techniques. The algorithm alternately optimizes user scheduling and association, UAV trajectory, and transmit power, with each iteration focusing on one block while keeping the others fixed. For the non-convex trajectory and power control problems, approximate convex optimization techniques are applied. Additionally, a low-complexity initialization scheme based on simple circular trajectories and circle packing is proposed to speed up convergence. Numerical results demonstrate significant throughput gains compared to static UAV or other benchmark schemes, highlighting the effectiveness of the proposed design.