Abstract: Rapeseed is one of the oil crops with the greatest expansion planting potential in China. Besides rice, rapeseed is also one of the most suitable crops for aerial seeding. The current aerial seeding system has been widely used in recent years, such as the spinner spreader by DJI, and the screw spreader by XAG. However, the seeds are scattered and disorderly after landing. The unstable working width cannot fully meet the requirement in the small and medium-sized fields. In an air-conveying rapeseed aerial seeding device, the migration trajectory of the seeds is easily disturbed during the process of leaving the end of the seeding tube to the soil, resulting in non-uniform seeds distribution. It is a high demand for an accurate system to save the seeds, lighting and ventilation for crop growth during weeding and cultivating in the fields. In this study, a systematic process analysis and component optimization were performed on the air-conveying rapeseed aerial seeding device. A kinematics model was also established for the seeds in the tube during the seeding process. DEM-CFD(Discrete Element Method-Computational Fluid Dynamics) gas-solid coupling simulation was implemented to determine the movement trajectory and velocity field of the seeds in the conveying airflow and the UAV rotor. A high-speed camera system was then selected to verify the simulation. After that, optimization was conducted for the structure and working parameters of the seeding device. Three seeding schemes were proposed during this time, including with/without the gas-solid separator, and the airflow accelerator. A series of performance tests were also carried out to determine the optimal structure, including the high-speed photography bench, the ground and real field aerial seeding test. The hybrid orthogonal experiment showed that the main and secondary order of the influencing factors was the conveying air velocity, the length of the seeding tube, and the inner diameter of the seeding tube, particularly for the speed change of the seeds in the vertical seeding tube. Furthermore, there was an extremely significant influence on the speed of rapeseed in the seeding tube (P<0.01). The inner diameter of the seeding tube presented a generally significant effect on the speed of rapeseed in the seeding tube (0.05≤P<0.1). The high-speed photography and ground mud box test showed that the conveying airflow speed increased significantly, as the original gas-solid separator was changed to a seed accelerator. Specifically, the airflow velocity was 4.5 m/s at the end of the seeding pipe, and the vertical displacement was 30 cm. The average horizontal displacement was 4 cm for the seed movement, and the width of the seed landing strip was 5.2 cm in the mud box at 30 cm directly below the seeding tube. The average horizontal displacement was reduced by 3.1 cm in this case, compared with the gas-solid separator. The width was reduced by 7.9 cm, which was basically consistent with the simulation. The field experiments showed that the outstanding germination and emergence of rapeseed were achieved, including the formed strips, and the row spacing, according to the optimized seeding scheme using seed acceleration. The improved system can fully meet the design requirements of rapeseed in the strips.