To address the issues of poor timeliness and delayed feedback in traditional wheat milling processes for manual particle size detection, a wheat milling online particle size detection device has been designed. This paper focuses on the design optimization of the key feeding component in the device, which affects the accuracy of particle size detection. The feeding component adopts shaftless screw blades. Through theoretical analysis, the main parameter ranges affecting the throughput capacity of the shaftless screw conveyor are determined. A Box-Behnken experiment is designed to obtain the optimal parameter combination for each factor: outer diameter of screw blades 23.8 mm, inner diameter of screw blades 6.4 mm, pitch 11.2 mm, and blade rotation speed 288.9 r/min. Simulation and test stand experiments are conducted using the above parameter combination. The simulation results show that the average throughput capacity of the feeding component is 2.85 kg/h, while the average throughput capacity of the test stand experiment is 2.84 kg/h, with a coefficient of variation of uniformity of 1.33%. These results indicate that the above parameter combination meets the design requirements for the feeding component in the device.