A Gas–solid two-phase flow coupling simulation is widely used to study the working process of pneumatic seed dischargers. Due to the demand for deterministic particle orbit numerical calculation models, seeds are mostly modeled using the particle aggregation method, where the seed model is formed through particle aggregation bonding without overlapping. The smaller the radius and the more filled ball particles used in this method, the closer they resemble the real morphology of the seed. However, this results in the over-consumption of simulation computational resources and simulation time growth. In this study, we used wheat seeds as the research object, studied the effect of seed models with different filled ball radii on the kinetic response characteristics between the particles, and searched for the optimal number of filled ball particles for the seed model. With the help of three-dimensional scanning and inverse fitting methods to obtain the seed profile, we used different radii (0.2 mm, 0.24 mm, 0.28 mm, 0.32 mm, 0.36 mm, and 0.4 mm) to fill the ball particles, and formed a wheat particle bonding model for a gas–solid coupling simulation. We used a combination of real tests and simulation measurements of bottomless cylinder-lifting and slip-stacking. The interspecies static and dynamic friction factors in seed models with different radii of filled spherical particles were first calibrated using the angle of repose as an index. Then, the parameters were verified using bottomless cylinder lifting and slip stacking tests, which used the coefficient of variation for the simulation test’s angle of repose as an index. Our results show that the smaller the radius of the filled ball, the closer the simulation results were to the real value. Validation was conducted using a gas–solid coupling simulation of an air-blown wheat seed discharger, with the seed filling rate as an index. Our results showed that the simulation length and simulation accuracy were optimal when the radius of the filling particle was 0.32 mm.