The improvement of the sugarcane yield in the geographic valley of the Cauca river in Colombia is the main goal of numerous experiments performed in the sugar sector of the country. The early estimation of sugarcane yield allows to identify the factors that affect the performance of a crop, accessing to act quickly to counter those effects. The use of RPAS equipped with modified cameras that allows to obtain images with information in ranges of wavelengths different to the visible spectrum has been in a peak for its fast, economic and easy manipulation and obtaining of support data for the precisión farming. The objective of this study was to generate a sugarcane yield estimation model based on variables obtained of images of a conventinal camera (RGB) and a modified camera (Red Edge) that allows to estimate the crop yield. There were made two flights over a sugarcane crop corresponding to the variety CC93- 4418 to obtain a mosaic from the modified camera and a Digital Heights Model – DHM from the conventional camera. With the mosaic were calculated five vegetation ńdices, of which was extracted their value for 54 plots making an average by index and were compared with the yield by means of linear regressions. The five indices obtained R2 over 0.7, being the GNDVIRe the one with the best R2 with a value of 0,787. The DHM was obtained from the Digital Elevation Model – DEM, generated by the Agisoft PhotoScan software, by subtracting the average elevation value of multiple ground points adjacent to the crop; from which was extracted the average value by plot, obtaining a R2 of 0.479 in relation with the height of the crop measured at 6 months of age and a R2 of 0.438 in relation with the crop yield. The good relatinship of the vegetation ‘d́ices and the DHM with the crop yield allowed to generate a model from a multiple linear regression of both variables, reaching a R2 of 0.808. The model was validated on 18 plots of an adjacent crop, reaching a R2 of 0.921.