Unmanned aerial vehicles (UAVs) are being used for rice seeding in agriculture. A new method using image recognition technology was developed to evaluate rice sowing performance. By optimizing the design and operation of the UAV-based seeding device, better seeding results were achieved. The size parameters of the guide tube were optimized for the best seeding effect. The seed discharge rate and rotational speed of the seed metering wheel were found to be crucial factors. Adjusting the friction wheel rotation speed at different heights improved seeding. By testing flying speed and seed metering wheel speed, better seeding uniformity was achieved. In a five-row seeding experiment, the width of seed rows and seeding uniformity were measured, providing valuable insights for improving UAV rice seeding.