Milk fat is an important nutrient for young animals and humans, and its composition [fatty acids (FA)] may vary broadly. The objective of this study was to describe the development of global FA models for high-throughput Fourier transform infrared milk analysers. The developed models for major FA, FA grouped by chain length, by degree of unsaturation and by their origin, showed good accuracy, repeatability and correlation to gas chromatography. Moderate performance was obtained for low concentration FA (short chain, polyunsaturated and trans). Possible practical applications could be to enhance dairy products, optimise milk processing and/or optimise dairy cow performance through improved feeding and management.