Aims: Insulin resistance underlies the etiology of both type 2 diabetes and gestational diabetes. In pregnancy, insulin resistance is also associated with an unfavorable metabolic programming of the fetus, potentially contributing to a higher risk of obesity and type 2 diabetes in the offspring. To assess insulin sensitivity, several methods based on glucose and insulin levels during a 75-g oral glucose tolerance test (OGTT) exist. It is unclear how they perform during pregnancy, where physiologically altered metabolism could introduce a bias. Methods: In a cohort comprising 476 non-diabetic subjects undergoing OGTT and hyperinsulinemic-euglycemic clamp (HEC), we used cross-validation to develop an insulin sensitivity index also based on non-esterified fatty acids (NEFA) that could be more robust during pregnancy (NEFA-index). We tested commonly used OGTT-based indexes and the NEFA-index in a different cohort of 42 women during pregnancy and 1 year after delivery. Results: The Matsuda and OGIS index failed to detect lower insulin sensitivity during pregnancy as compared to the follow-up OGTT 1 year after delivery (p > 0.09). The new NEFA-index incorporating BMI, plasma insulin and NEFA, but not glucose, clearly indicated lower insulin sensitivity during pregnancy (p < 0.0001). In the non-pregnant cohort, this NEFA-index correlated well with the gold-standard HEC-based insulin sensitivity index, and outperformed other tested indexes for the prediction of HEC-measured insulin resistance. Conclusions: This insulin/NEFA-based approach is feasible, robust, and could be consistently used to estimate insulin sensitivity also during pregnancy.