In order to solve the problem of high reseeding, miss seeding and instability issue in the vegetable tray precision seeder, the seedling performance was detected in the tray precision seeder based on machine vision technology. The performance detective device was setting in the tray precision seeder. Double camera took tray images one by one after the tray was detected by photosensor, and then these images were transferred to the computer in real time. The images were processed and analyzed by the vision algorithm software, the seeds could be seen in the tray and the results and photos were displayed on the screen together. The test showed that the reseeding detection accuracy was 98.94% and miss-seeding detection accuracy was 99.33%.