Abstract: Drought is an important issue seriously affects the normal sowing and seed germination of dryland crops such as corn and cotton. The fluid seeding technology suspends seeds in a mixture named as seed-liquid of water and high absorbent polymer (HSP), which absorbs water when high in moisture and slowly releases water when lacking, to provide a favorable condition for the germination and emergence of the crop. This technology has multiple advantages such as drought resistance, water conservation, ensuring emergence rate, reduced seed damage rate, and increased yield. Uniform distribution of seed in the liquid is essential for fluid seeding quality, and pneumatic agitation is a good way to form the flow field for the suspension of the seed. However, the flow field is significant affected by the structure of the seed-liquid tank. Four seed-liquid tanks of different shapes—square (SQ), triangular type (TRA), quasi wedge (QW), and quasi wedge-arc (QWA) were designed for the developed pneumatic-agitation fluid seeder in this study. During operation, air is supplied through the air inlet at the bottom of the seed-liquid tank. A relationship model between the air consumption for pneumatic agitation and the structural parameters of the tank is established. The effect of the vertical section shape of the seed-liquid tank on the flow field of seed-liquid under different air inlet speeds was studied based on CFD simulation. The results show that the SQ seed-liquid tank has dead space in the circulation, which may cause seed deposition; the TRA seed-liquid tank does not have an overall circulation which is conducive to the uniform distribution of seeds, and the seed-liquid flow field in the middle and lower parts of the tank is unstable; the QW seed-liquid tank has multiple small vortices affecting the flow field of the seed liquid; and the QWA seed-liquid tank exhibits an optimal performance at all tested air inlet speeds with respect to seed-liquid circulation patterns, flow field stability, and the reduction of circulation dead space; when the inlet air speed keeps constant, the ratio of the maximum and minimum values of the seed-liquid flow speed varies within a range of 0.04, with minimal fluctuations and optimal stability in flow speed. The validation experiment is conducted on the QWA seed-liquid tank to verify the results of simulation tests, and it indicated that the stability of the seed-liquid flow field was optimal at 4 m/s of the inlet air speed. The error between the simulated value and the test results of the seed-liquid flow speed at the ideal seed outlet area in the seed-liquid tank is 9.3 %, which verified the reliability of the simplified gas-liquid two-phase flow model. The pneumatic agitation test was conducted at varying air flow rates to verify theoretical calculations through observing the distribution of seeds in the seed-liquid tank under different conditions. The test results show that the seed deposition decreases and distributes relatively uniformly in the liquid at inlet air flow rates of 1.5 and 2.0 m³/h, with corresponding air consumption values of 6.67×10^-3 and 8.89×10^-3 m³, which were consistent with the predicted value. The results may provide a foundation for the design of fluid seeding seed-liquid tanks.