During aerial photography, drones are prone to changes in posture due to external factors such as airflow and wind speed, which can affect the stability of the aerial footage. In order to improve the quality of drone aerial photography and achieve stable shooting of the target area, a robust attitude control method for drone aerial photography based on real-time state feedback is proposed. Consider aerial photography tasks and environments, and set unmanned aerial vehicle (UAV) attitude control objectives through UAV aerial photography route planning and attitude calculation. The use of real-time state feedback technology to determine the initial value of drone attitude facilitates faster response to changes in drone attitude. Set control constraints from both yaw and pitch angles to achieve all-round attitude control of the drone during aerial photography. Selecting the wings and power system of the drone as the object of robust attitude control for aerial photography, taking into account the flight environment and control calculation results of the drone, dynamically generating robust control commands for the drone"s aerial photography attitude, and achieving robust control of the drone"s aerial photography attitude. The conclusion drawn from the experiment is that the real-time feedback deviation of the state is less than 0.2 °, which meets the requirements of robust attitude control operation for drone aerial photography. The attitude angle execution deviation coefficient is reduced by 0.685 °, and the attitude angle execution deviation coefficient is significantly reduced.