Drought is currently one of the main causes of grain yield wheat losses in the world. Therefore, it is necessary that breeding programs characterize and select quickly and efficiently genotypes tolerant to water stress. A good alternative is the use of multispectral images, which allow recording the evolution in space and time, of the development of the crops, through biophysical parameters derived from that images. This work began season 2018/19, and will continue by three periods. Our aim is evaluate the use of NDVI and RGB to provide accurate and non-destructive estimates of the physiological traits that determine the performance of a genotype in each environment and finally its grain yield potential. To this end, a set of 16 spring bread wheat genotypes, with different degrees of dry tolerance in grain filling stage, was established in two Chilean highly contrasting environments, under water stress, Cauquenes (35°58′ S, 72°17′ W; 177m.a.s.l.), and under full irrigation and moderate water stress, SantaRosa (36°32′ S, 71°55′ W; 220 m.a.s.l.). By combining the climatological data of each zone, RGB and NDVI we want to determine the difference between the growth rate of genotypes, biomass, plant height, duration of the vegetative and fill grains stages and determine the effect of water stress in the NDVI, associating all this with the genotype´s grain yield. Our results are preliminary, but are coherent and show clear trends. At the end of three seasons we hope validate and integrate the use of these optical tools in the wheat breeding program in Chile, to increase the accuracy of selection in the field genotypes tolerant to water stress during grain filling period.