In colder parts of the North America, red clover (Trifolium pratense) is frostseed-established into pastures to increase forage legume density in the pasture. Improving this trait through breeding requires approximately 300 seed or more per genetic unit being evaluated and can only be field tested once a year, making selection for this trait difficult. Having a rapid and easy pre-screening method of red clover lines and germplasm would enhance the process. This study examines using near infrared spectroscopy (NIRS) scans of red clover seed and accompanying field data on red clover halfsib families to develop NIRS calibration equations for predicting red clover frost-seedability. Calibration equations were developed based on 40 halfsib families, whose frostseeded-establishment was measured under field conditions. Equations developed had r-square values of 0.5. These equations were used to predict the frostseeded establishment of independent seed lots. These same seed lots were evaluated for frost seeded establishment in the field. Predicted and actual frostseed-establishment was compared.