This dataset contains data used to generate results presented in the paper “Spectral differentiation of sugarcane from weeds", published in Biosystems Engineering Journal, 2020. Spectral data were collected with Fieldspec 4 Std-Res spectroradiometer (Analytical Spectral Devices Inc., Boulder, CO, USA) from sugarcane and weed leaves in two places in São Paulo State, Brazil. The data spreadsheet is comprised by the following columns: 1) Identification; 2) Two collection places - UFSCAR and FEAGRI; 3) A data classification with two classes - DANINHA for weeds and CANA for sugarcane leaves; 4) A data classification with three classes - sugarcane leaves (CANA) and weeds divided on monocotyledons (i.e. grass weeds - MONOCOTILEDONEA) and dicotyledons (i.e. broadleaf weeds - DICOTILEDONEA); 5) the following columns are the absorbance values starting from 450nm until 2400nm, with each value being the average spectra. Nine spectra per leaf sample, resulting from the integration of ten consecutive readings, were taken from the leaves' middle region and acquired immediately after field collection.