Our objective was to develop and validate a Fourier transform mid-IR-based milk analysis method to estimate blood NEFA concentrations in lactating dairy cows. High blood NEFA indicates that a cow is mobilizing body fat and increases the risk of metabolic disorders. Milk and blood samples were collected from 60 lactating Holsteins once per week for the first 3 weeks of lactation. Cows were milked 3 times per day. Within + or – one milking of the time of blood collection, a milk sample was analyzed using a Delta Instruments (model FTA) mid-IR milk analyzer. A Wako NEFA HR test kit was used as an in vitro enzymatic colorimetric method for the quantitation of NEFA in blood serum and these values were used as reference values for development of a partial least squares (PLS) regression model to predict blood NEFA from the mid-IR milk spectra. There are no NEFA in milk, so a model to predict blood NEFA from a milk sample uses differences in the milk spectra from sample to sample that are correlated with changes in blood NEFA. The final PLS model had 9 factors, used wavelengths in the following ranges (3000 to 2800, 1800 to 1700, 1585 to 1000 cm⁻¹) with a standard error of cross validation of 172 μEq/L. Validation milk and blood sample pairs (n = 53) were collected from Holstein cows from a different herd. The mean value for the blood reference test was 713 μEq/L of serum and the mean value for the milk based blood NEFA prediction was 703 μEq/L of serum with a standard deviation of the difference (SDD) of 218 μEq/L for the 53 validation samples. Blood NEFA measured on blood is a snapshot of the NEFA concentration at an instant in time, while blood NEFA predicted from milk analysis represents a time average for the total time between milkings. The FTIR milk analysis to estimate blood NEFA is rapid (about 10 s), done simultaneously with all other milk component measures, and uses no reagents. This approach could be useful for rapid evaluation of risks of ketosis and reproductive disorders.