The discrimination of scented rice is a major stake for the producers. There is however not any fast and reliable method for discriminating scented from non scented rice. In this perspective, we have developed tools based on the multivariate discriminated analysis of the spectral data obtained by Near InfraRed Spectroscopy (NIRS) or Mass Spectrometry (MS). Spectral data of whole rice grains were registered between 800 and 2500 nm. A set of 117 samples including 65 scented and 52 non scented rice was analysed. A model set comprising 90% of these samples was used to develop the discrimination model using the SIMCA procedure. The validation set was then predicted with the model with a success rate of 84%. Rice volatile compounds were extracted by SPME and directly injected in MS apparatus to obtain a mass spectral fingerprint of the volatile fraction. A set of 114 samples including 57 scented and 57 non scented samples was analyzed. A discrimination model was developed using SIMCA procedure on a model set of 102 samples. The validation set (12 samples) was predicted with the model with a success rate of 91,6%. The discriminated model was proved to be linked to some specific mass fragments resulting from 2-acetyl-1-pyrroline, the main aroma compound of scented rice. Both methods proved their ability to rapidly discriminate scented rice and practical automated procedures are under development.