The Cirad/CIAT collaborative project has developed a breeding program based on the recurrent selection (RS) for an eco-efficiency use of water. Selection was primarily based on yield and, in the course of the selection process, screening methods were refined with additional secondary relevant traits (morpho-pheno-physiological traits). The infrared (IR) thermography measures the plant canopy temperature, which gives an indication of the plant water status through leaf surface cooling capacity by transpiration along environmental conditions. We developed large scale phenotyping methods, adapting the IR technology to field conditions, to evaluate response to drought among synthetic populations. An experiment was conducted during the 2008/2009 dry season in the Santa Rosa CIAT research station in Villavicencio (Colombia) to adjust the phenotyping method based on IR thermography and to evaluate 400 S1 lines and 1000 S0 plants derived from synthetic populations. This evaluation method allowed us to identify interesting progenitors to be integrated in the population improvement scheme with the goal to develop and diffuse improved material, populations or lines, with increased resilience to drought for various rainfed ecosystems in Latin America and the Caribbean (LAC).