Evidence suggests that the source of dietary protein may have an impact on insulin resistance, but no studies have explored it in pregnant populations. In this study, we combined a population study and an animal experiment to explore this effect. The population study was conducted with data from NHANES. Multiple linear regression was used to observe the association of protein intake with outcomes, including fasting glucose (GLU), insulin (INS), and HOMA-IR. In the animal experiment, 36 pregnant SD rats in three groups were orally administered 100% animal protein, 50% animal protein and 50% plant protein, or 100% plant protein, respectively. The intervention continued throughout the whole pregnancy. On day 19.5, maternal plasma was collected after overnight fasting, and metabolomics was performed using UPLC-MS. We found plant protein intake was negatively correlated with INS and HOMA-IR in the whole population. During the third trimester, a similar correlation was also observed. The animal experiment also presented the same result. In metabolomic analysis, changes in various metabolites and related pathways including FoxO and mTOR signaling pathways were observed. In conclusion, we found a negative association between dietary plant protein intake and maternal insulin resistance during pregnancy. Changes in some active substances and related metabolic pathways may play an important role.