Cannabinoids in hemp plant are strictly located in the inflorescence, in particular in top side of the plant while a lower amount may be found in the leaves. As a consequence, the lower amount of cannabinoids which can be recovered in the hemp flours is obtained from seeds, thus from contamination during harvesting procedures. To this aim, a screening test for the real time detection of cannabinoids in hemp flour was developed by a miniaturized analytical platform based on the MicroNIR spectrometer. Chemometrics was used to develop models of prediction to identify the cannabinoids and simultaneously to quantify the residual amounts in order to accomplish specific regulatory legislation according to the country. In particular, Partial Least Square-Discriminant Analysis (PLS-DA) and Partial Least Square regression (PLSr) were applied and results demonstrated that MicroNIR/Chemometric platform permit to differentiate hemp flours according to the presence of CBD, THC and CBG. In addition, for each cannabinoid, a quantification method was developed in the range 0.001-0.1 %ww, and their performances were evaluated by comparing results to those obtained by the reference procedures. Processing of the real samples provided a suitable correlation and confirmed the capabilities of this innovative platform to be used for the monitoring of the residual content of cannabinoids in hemp flours.