Mung bean is the main edible bean in China, which has high economic and edible value. However, the current level of mechanized planting of mung bean in China is still low, and sowing or seeding is used. The waste of improved varieties is serious, and the yield and planting efficiency are low, which seriously restricts the development of mung bean industry. In view of the characteristics of scattered mung bean planting and small plot planting in China, and the agronomic requirements of 2~3 seeds per hole, a mung bean precision seed-metering device with disturbance for promoting seed filling was designed. The seed metering device used concave holes to carry seeds, the seed guide groove to promote seed filling, and the brush to clear seeds and protect seeds. The mechanical relationship of disturbance filling promotion was analyzed, and the design method of parameters of seed guide groove, seed carrying hole and brush was determined. The discrete element software EDEM simulation optimization method was used to establish the discrete particle model of mung bean seeds and the experiment was employed through the method of orthogonal test of three factors and three levels with the structural dimension parameters of the seed carrying holes as experimental factors. Three test parameters, i.e. length of the seed hole, width of the seed hole and the depth of seed hole were chosen as the influence factors of test experiment. The best work parameters were as below: the length was 10.5mm, the width was 6.5mm and the depth was 5mm. The results showed that the optimal seed charging rate was 96.05%, the qualified seed charging rate was 1.64% and the missed charging rate was 1.57%, which were consistent with the simulation optimization results. Then, a speed single factor test was carried out. The results showed that when the working speed was within 8km/h, the qualified index was higher than 90%, the missing index were not more than 5%, and the multiple index were not more than 2%. At last, the adaptation tests were carried out. The results showed that qualified index was more than 95%, the missing index was not more than 5%, and the multiple index was not more than 3%, which could completely meet the requirement of the technical specifications of quality evaluation for mung beans metering device in China. The research provided a reference for further improving quality and optimizing the parameter for mung bean metering device.