Abstract: Abstract: A pollinator with high-clearance chassis has an excellent performance in hybrid rice beyond artificial pollination with high labor intensity and low efficiency. However, the lack of pollination often occurs when the pollinator is separated from the male parent panicle layer, due to the large size of the left/right trusses swinging in the vertical direction, particularly in the uneven bottom soil. In this study, a pipeline-airflow pollinator was designed to embed panicle layer for the better pollination of hybrid rice. A lightweight vertical pollination tube was adopted, where the effective working length was 60mm. As such, the panicle layer in the local paternal area was effected by the pollination wind field always even if high-clearance chassis shaken when the pollination machine was walking in the field. In addition, the key components were optimized to obtain a uniform airflow field and a suitable velocity of airflow in a pollinator. Firstly, the flow divider was simulated to evaluate the distribution of airflow. Then the circular-hole jets were adopted to adjust the pollination airflow field under the different diameters of the nozzle. Simulation experiments were also conducted on pollination tubes with various nozzle diameters. The simulation results showed that there was no significant effect of inlet Inlet air volume on the airflow distribution of divider in the working range (P> 0.05). An optimal splitter was selected when the average coefficient of variation was only 1.14% at the inlet total pressure in three pollination tubes, indicating a better uniform airflow field. The air velocity along the pollination boundary was 1.2m/s just above the suspended velocity of pollen, when the diameter of the pollination tube nozzle was 12 mm. A prototype was also trial-produced for the field tests. The experiment results showed the simulation data was basically consistent with the actual one. The relative error of airflow between simulation and measurement was ±0.3 m/s. The coefficient of variation was only 1.12% for the axial airflow velocity of the nozzle in three pollination pipes, when the diameter of the nozzle was 12 mm, indicating suitable for the design requirements of the uniform wind field. The distance of the pollination tube from the middle of the panicle layer was less than half of the effective working length of the pollination tube when the speed of the pollination machine was 1.5m/s, indicating an effective range of pollination wind field. The pollination experiments were conducted in the paddy field under the condition that the walking speed of the pollination machine was 1.5m/s and the average axis velocity of airflow at the pollination tube nozzle was 34.2 m/s. A total of 80 acquisition points were divided into 5 lines on average in the female compartment area against the male parent. The field experiments showed that the average number of grains in the unit visual field was 8.35 (hybrid rice seed production agronomically requires at least 3 grains of pollen), of which the average number of pollen grains more than 3 grains accounted for 96.02%, indicating the minimum requirement of pollen quantity for hybrid rice seeding and pollination in local areas. This finding can provide a sound reference for the production of hybrid rice seeds in mechanized agriculture.