The world production of dates has more than doubled over the last two decades. Presently, the industrial sorting of dates is a difficult operation, affected by a high level of losses. As a consequence, the objective of this article is to study the variety discrimination capabilities of Fourier Transform-Near Infrared (FT-NIR) reflectance spectroscopy for post-harvest sorting purposes. The same methodology was tested on six commercial date varieties named Estamaran, Kabkab, Mazafati, Piarom, Deglet nour and Zahedi. For each cultivar, 100 fruits were analysed using the FT-NIR spectrometer, subsequently characterised by destructive measurements. A multivariate analysis of the FT-NIR spectra appears a feasible tool for cultivar discrimination and even to estimate the soluble solid content (SSC) and dry matter (DM) of dates at their optimal harvesting conditions. FT-NIR, as a non-destructive techniques, could be a suitable way for on-line screening of different dates varieties.