Abstract: In order to select high-quality and high-protein soybean suitable for planting in northwest of Shandong, fuzzy membership function and principal component analysis were used to evaluate the production performance of 10 high-protein soybean varieties. The results showed that the coefficient of variation of branch number, 100-seed weight, bottom pod height, yield and plant height were larger, but the coefficient of variation of main stem node number, pod number per plant and seed number per plant were smaller. The number of branches was negatively correlated with plant height and node number of main stem, and positively correlated with the pods number of per plant. There were significant positive correlations between pod number and grain number of per plant, as well as between 100-grain weight and yield. Therefore, varieties with more branches, pods and grains per plant, 100-grain weight should be appropriately selected when breeding high-yield varieties in this region. According to the membership function and principal component analysis, Jidou 12, Qihuang 34 and Hedou 37 have high comprehensive scores and high yield, which has strong adaptability and can be used as the preferred high-protein variety for planting in northwest of Shandong.