The aim of this study was to assess near-infrared reflectance (NIR) spectroscopy as a tool for determining sensory and texture characteristics of beef. Chemical, instrumental, texture and sensory characteristics were predicted by near-infrared reflectance spectroscopy carried out on longissimus dorsi muscle samples from 190 young bulls. The use of first derivative gave best predictions together with NIR spectra, except for myoglobin and water holding capacity, which had an R(2) of prediction of 0.91 and 0.82, respectively, using visible and NIR spectra. Tenderness was the best-predicted variable (R(2)=0.98) demonstrating the potential of NIR spectroscopy in the prediction of sensory variables. Chemical composition variables and Warner-Bratzler shear force were predicted with an R(2) of prediction of around 0.7, but protein was not predicted with accuracy.