Shrubs are an important source of food for domestic livestock and wildlife in semiarid areas. This is particularly true during dry periods when forage from natural pastures is scarce and its quality is low. Cytisus multiflorus is a leguminous shrub which is endemic to the Northwestern Iberian peninsula. This species occurs in poor soils and in degraded or marginal areas. The objective of this study was to evaluate the potential of near infrared reflectance spectroscopy (NIRS) to determine quality attributes of Cytisus multiflorus as a forage feed. NIRS calibrations were developed for crude protein, neutral detergent fibre (NDF), acid detergent fibre (ADF), acid detergent lignin, cellulose, hemicellulose, ash, and organic dry matter digestibility (OMD). The sample population used included different plant parts and plants of different ages, and was collected in the Spanish province of Salamanca. Calibration models were accurate for the prediction of crude protein (r2 = 0.99; Standard error of prediction (SEP) = 3.8 g kg-1), OMD (r2 = 0.98; SEP = 15.7 g kg-1), NDF (r2 = 0.98; SEP = 13.3 g kg-1), ADF (r2 = 0.94; SEP = 19.1 g kg-1), cellulose (r2 = 0.96; SEP = 12.2 g kg-1), and ash (r 2 = 0.91; SEP = 1.1 g kg-1.1 g kg-1). The results of the lignin and hemicellulose calibrations had lower accuracy (r 2 = 0.76; SEP = 7.0 g kg-1 for lignin and r2 = 0.88; SEP = 13.6 g kg-1 for hemicellulose). The results showed that NIR spectroscopy can be used for the rapid and accurate prediction of quality attributes in samples of the mediterranean shrub Cytisus multiflorus. This work was supported by the “Junta de Castilla y León” from Spain (Project No. CSI 6/01).