Currently, quadrotor UAVs are widely used in different fields. However, UAVs are often hindered by obstacles during route travelling, which leads to the deviation of UAVs from the intended route. For this reason, the study optimises and analyses the trajectory of the UAV, and proposes a new model combining the optimised gradient descent algorithm and the A* algorithm, which optimises the A* algorithm through the gradient descent algorithm. After experimental testing the new model in the UAV trajectory optimisation effect has been significantly improved, the new model than the traditional algorithm model displacement deviation reduced by 0.34m, angular deviation reduced by 0.32rad, obstacle avoidance efficiency has been improved, it can be seen that the use of the new model can effectively improve the efficiency of the flight of the UAV, to avoid the UAV airway stays, to improve the safety of the route and the research uses the new model can effectively improve the UAV trajectory optimisation effect, which is a good reference value for future quadcopter UAV trajectory optimisation performance improvement.