Abstract: Abstract:  In the case of non-uniform user distribution, a unmanned aerial vehicle integrated sensing and communication system design based on non-orthogonal multiple access (NOMA) is conducted. After clustering multiple randomly distributed users, the research focuses on drone trajectory design, beamforming, and power allocation, and establishes a joint optimization problem. The objective is to maximize the minimum user rate under the premise of ensuring sensing performance via radar echoes. Due to the coupling of multiple variables in the established optimization problem, which belongs to a non-convex problem, after determining the drone's flight trajectory according to the clustered user conditions, the block coordinate descent method is employed to decompose the original problem into two subproblems. Then, by introducing slack variables, using first-order Taylor expansions, and applying successive convex approximation, the subproblems are transformed and solved. Finally, the approximate optimal solution to the original problem is obtained through alternating iterative optimization of the subproblems. Simulation results demonstrate that the proposed algorithm exhibits good convergence and can effectively improve the system's average throughput.