Abstract: Aiming at the increase of working speed, the seed filling of air suction precision maize seed metering device is not good, which leads to the serious missing seeding and high demand for wind pressure. A new type of air suction high speed precision maize seed metering device with assistant seed filling plate was designed. The seed metering device could use the type hole embossement to disturb the seeds in the filling area, and play the assistant filling effect. In the clearing and carrying areas, the seeds could be held to reduce the possible falling from the seed plate; the seeds in the seed unloading area could be planted evenly under the power of the type hole luge. The design of type holes must ensure adequate seed filling, effective seed cleaning and smooth seed metering. To realize the assistant filling of the air suction seed plate by using the type hole embossment to disturb and support the seeds, the working area of seed metering device was analyzed, and the key structural parameters of seed metering device were calculated. It is necessary to increase the seeds disturbance with the assistant filling. The rotation of the seed plate is faster under the high speed condition, the contact time of seeds with type hole is short, the seeds are not easy to be adsorbed, and the disturbance of the seed is beneficial to the contact of the single seed with type hole, so the disturbance performance of the seed plate will directly affect the seed filling performance of the seed plate. The simulation analysis of seeds disturbance was carried out by using EDEM (engineering discrete element method) software. A three-dimensional model was established in SolidWorks based on the shape characteristics of 1000 seeds, and then a multi spherical composite filling method was used in EDEM to make it stacked into particles close to the outline of the seed, as the simulated particle body of the maize seed. The average distance of the two seeds, the average speed of the seeds and the mean velocities of the seeds under different angular velocities were evaluated as three indexes. The results showed that the designed seed plate could effectively enhance the seeds dispersion and play the role of disturbing the seeds. And then the laboratory tests were carried out. The Zheng Dan 958 was selected, its weight was 351 g, and the water content was 12.8%. The performance detector of the seed metering device, which was developed by China Agricultural University was selected, and the theoretical spacing of the plant was set to 25 cm. The 4 kinds of seed plates used in the simulation were tried and made. The three factors orthogonal test is carried out with the factors of speed, wind pressure and type of seed plate. The key factors that affect the qualified index, multiple index and missing index were determined by the difference analysis and analysis of variance. The regression analysis was taken to select the better results, and the regression equation is obtained. The test results showed that the designed seed plate working index was better than other kinds of seed plate. Using the design expert 8.0.6 software to process the data, the regression model between the performance index of the seed metering device and the speed, and the wind pressure were established. Through the response surface analysis test, factors influence on the response index were analyzed, the multi-objective optimization method is adopted, and the qualified index, the missing index and the multiple index were used as the evaluation indexes. The response surface of the qualified index and missing index under different speed and wind pressure were obtained, and the order of influence of each factor were analyzed intuitively. The optimum operating parameters of the designed seed metering device at the speed 8, 10, 12, 14 km/h are obtained. The results showed that the design plate could effectively improve the intensity of the seeds disturbance and reduce the missing index of high speed operation, and under the operating speed of 8 to 14 km/h. The negative pressure value is between -3.21−-3.71 kPa, the qualified index was more than 91.3%, the missing index was less than 5.1% and the multiple index was less than 3.6%.