Abstract: Pulsed-jet orderly seeding is one of the most promising technologies using unmanned aerial vehicles (UAVs). A seeding pump is often used to pressure the seed-liquid mixture (liquid and germinated seeds), thereby increasing the ejection velocity of rice seeds to suppress the influence of rotor airflow on seed falling trajectories. Orderly row-and-hole seeding can be expected for rice planting. However, conventional seeding pumps cannot fully meet the requirements of pulsed-jet orderly seeding, in terms of flow rate, seed damage rate, solid passing rate, and discharge pressure. In this study, an optimal pump configuration was selected in a pulsed-jet pump for rice seed metering. Comparative experiments were conducted on three types of seeding pumps: an axial flow pump, a helical impeller pump, and a semi-open impeller pump. Flow rate, seed damage rate, solid passing rate, and discharge pressure were taken as evaluation indicators, with weights determined by the Analytic Hierarchy Process (AHP) at 16.76%, 23.11%, 48.41%, and 11.72%, respectively. The experimental data were also normalized for dimensionless comparison. The comparative test results showed that for flow rate, material passing rate and ejection pressure, the performance followed the order: semi-open impeller pump > axial flow pump > spiral impeller pump. While the seed damage rate followed the order: spiral impeller pump＜semi-open impeller pump＜axial flow pump. The evaluation indices were 0.464, 0.231, and 0.985, respectively, for the axial flow pump, spiral impeller pump, and semi-open impeller pump. In terms of its superior seeding performance, the semi-open impeller pump was selected as the seeding pump configuration for the UAV pulsed-jet rice orderly seeding, providing for subsequent structural optimization and field prototype testing. A Computational Fluid Dynamics (CFD) model was established to optimize the structural parameters of the semi-open impeller pump. The area-weighted average pressures were monitored at the pump inlet and outlet, together with the impeller torque. Hydraulic efficiency was used as an evaluation metric. A single-factor experiment was conducted to investigate the effects of key impeller parameters: number of blades, impeller diameter, and blade width on the hydraulic efficiency. On this basis, a three-factor, three-level Box-Behnken experiment was conducted to investigate the interaction of impeller structural parameters on pump performance. Flow rate and seed damage rate were taken as evaluation indicators. The results indicated that the number of blades, impeller diameter, blade width, and the interaction between the number of blades and blade width shared significant effects on both flow rate and seed damage rate. The significance of each parameter was ranked as the number of blades > impeller diameter > blade width. With maximum flow rate and seed damage rate below 0.8% set as optimization objectives, multivariate regression analysis showed that the optimal combination of parameters was determined to be 4 blades, an impeller diameter of 76 mm, and a blade width of 22 mm, which yielded a flow rate of 1.12 kg/s and a seed damage rate of 0.74%. A prototype semi-open impeller pump was manufactured with the optimal parameters. Field tests were conducted at the La Pu experimental site in Zengcheng District, Guangzhou Province, China. The average coefficient of variation of inter-hole spacing was 13.2%, the mean standard deviation of seed hole center offset distance was 2.0 cm, and the rice yield was 6731.03 kg/hm². Therefore, the UAV pulsed-jet rice orderly seeding can achieve orderly row-and-hole planting, indicating the high feasibility of the seeding pump in the field.