The amount and fatty acid (FA) composition of beef intramuscular fat (IMF) are key factors that influence technological and sensory quality, especially shelf-life (lipid and pigment oxidation) and flavour. Furthermore, consumers are interested in the fat composition of meat, as nutritional guidelines are recommending a lower saturated FA (SFA) intake due to its association with cardiovascular diseases. The amount and composition of ruminant IMF, which depends on factors such as the genetic origin of the animals, feeding regime, age or live weight, influences the final quality of the product, which also explains the increasing interest in defining the FA profile of meat. However, quantitative chemical techniques for the determination of FA involve extraction of total lipids and determination of FA methyl esters by gas chromatography, so that this procedure is costly, time-consuming and generates hazardous waste. The use of near infrared reflectance spectroscopy (NIR) is increasing in food analysis because it offers several advantages over conventional methods, giving fast, non-destructive, clean and cost effective measurements. Therefore, the aim of this study was to test the on-line estimation of the concentration of major individual FA (C16:0, C18:0 and C18:1) and main groups of FA (SFA, monounsaturated FA (MUFA) and polyunsaturated FA (PUFA)) of beef IMF using NIR, by direct application of a fibre-optic probe to the M. longissimus thoracis with no prior sample treatment.