The growing consumer demand for high-quality pork produced with high-quality products must be accompanied by monitoring to ensure their traceability. In a study to evaluate effects of genetic and nutritional factors on meat quality, we measured backfat and muscle tissue by near infrared spectroscopy (NIRS). This work aimed to develop (i) a model to differentiate pigs according to their feed composition and (ii) a model to predict the fatty-acid (FA) composition of pig meat from NIRS spectra. Spectra (n=160/tissue) of backfat and longissimus thoracis muscle were acquired from pigs of different genetic types ([LW×LD] x Piétrain; [LWxLD] x Duroc; Danbred x Duroc). In parallel, tissue FA composition was determined by gas chromatography. Intra-genotype, half of the animals received a control diet based on oilseed meal, and the other half a diet including faba bean and linseed selected and specifically treated, as the main sources of protein and n-3 polyunsaturated FA (PUFA), respectively. NIRS technology was able to differentiate pork samples according to the pig diet, as 100% of the backfat samples were correctly ranked during the validation step. The muscle matrix differentiated less well than the backfat (accuracy = 75%). In backfat, FA composition was predicted well for C18:3n-3, n-3 PUFA, n-6 PUFA and saturated FA (R² of cross-validation= 0.84, 0.83, 0.78 and 0.85, respectively, and an associated ratio of performance deviation greater than 2). In conclusion, NIRS is a rapid tool to differentiate pig tissue samples according to a specific diet and predict their FA composition, especially for backfat.