Qualified rate of the air suction seed metering device of 2BQ series corn seeder in seeding Duyu1, Longdan38, Xianyu335 and Xinxin6 was tested by using JSP–12 metering device. Regression prediction model and the BP neural network model was used to forecast the gas suction qualified rate of corn seed metering device. The results show that in the speed of 6.0–12 km/h, sowing Duyu1 the percent of pass is 86%–96% of seed, Longdan38 the percent of pass is 71%–94%, Xianyu335 the percent of pass is 79%–92% of seed, Xinxin6 the percent of pass is 78%–96%. The BP neural network model for gas suction qualified rate of corn seed metering device has good fitting capability and relatively high prediction accuracy.