In corn fields, weeds often appear in patches and applying herbicides uniformly across the field is not optimum. A better approach would be to target application to field areas where there is weed infestation (spot spraying). Spot spraying requires that weed infested areas are identified either prior to herbicide application or in real-time during application. For this paper, 3 years of experimental data on the use of UV-induced fluorescence for discriminating between corn and weeds were available. If discrimination is feasible, this technology could be used for real time identification of weed patches in corn fields. Plants were grown and measured under contrasting conditions. Plants were grown either under artificial lighting in a growth chamber or in a greenhouse under natural sunlight. Measurements were performed either in the greenhouse under natural sunlight or in a growth chamber under standardized lighting conditions. Several factors were considered: weed type (monocots or dicots), leaf side and position of measurement on the leaf. Plant fluorescence was induced around 328 nm and fluorescence measured between 400 and 755 nm. For discriminating between monocots (including corn) and dicots, a robust classifier can be constructed using partial least-square discriminant analysis. Such a model based on two large spectral bands of the emitted fluorescence (400420 nm and 470520 nm) provided robust inter-year results with classification error from 1.2% to 4.6%. Discriminating between corn (a monocot) and monocot weeds can be achieved but with a lower success rate. In cross-validation, best PLS-DA models based on four 22.5 nm bands resulted in classification error smaller than 7.4%. In prediction, the classification error was high (5.6% to 45.2%). A new normalization scheme was defined based on mean values for blue-green and chlorophyll fluorescence. With this scheme, discrimination of corn from monocot weeds was performed with a maximum error of 16.6%.