The AMMI model was used to analyze the multi-environment trials data of maize varieties in Hebei Province in 2017. The results showed that genotype effect（G）, environment effect（E） and genotype×environment interaction effect（G×E） reached extremely significant level, in which the environmental effect accounts for a maximum of 79.79% of the total variation, genotype and environment interaction account for 10.30%, and the genotype effect only accounts for 2.84%. The first principal component axis（PCA1）, the second principal component axis（PCA2） and the third principal component axis（PCA3） in the genotype-environment interaction effect accounted for 83.14% of the interaction information. Among the 12 tested varieties, G7（Hengyu1587）, G9（Wodan901） and G4（Wodan818）were high-yield and stable varieties. The control variety G12（Zhengdan958） was good stability with general yield.Among the 9 testing sites, E2（Gaocheng） had the best resolution, and E8（Yongnian） was a testing site with the poorest resolution.