The Senegal River basin flows along the borders of Senegal, Mauritania, Mali, and Guinea and provides irrigation for over 200,000 ha of irrigated rice. During the harmattan season (December-March) rice is not cultivated and fields are typically left as fallow. In this short window, durum wheat was successfully cultivated despite max daily temperatures above 31°C throughout the season. Field testing here at two sites over two seasons of a large panel of modern entries genotyped with 8,173 polymorphic SNPs allowed to identify 36 genomic regions involved in the control of various heat tolerance traits. Testing in Senegal during two seasons a second independent validation set of germplasm genotyped with 7,992 polymorphic SNPs confirmed ten genomic regions as being associated and not involved in the control of phenology. Finally, a previously published study using the same germplasm and exposing it to heat stress via the deployment of plastic tunnels at the time of flowering, further confirmed the importance of four of these genomic regions on chromosomes 1A, 5A, 2B, and 3B. A haplotype study confirmed that having the positive allele at Q.icd.Heat.002-1A, Q.icd.Heat.012-2B, and Q.icd.Heat.018-3B resulted in 10%-20% higher grain yield under heat stress. The sequence of the array markers underlying these regions (AX-95099328, AX-95184164, and AX-94932858, respectively) were converted into KASP and used to genotype a third independent set of ICARDA’s elites, also field tested in Senegal. These markers explained 4%-13% of the phenotypic variation for grain yield under heat and are hence considered as validated for MAS and the whole community can now use them for this scope. In parallel, a genomic selection pipeline was built to combine the best minor alleles with the major one described above. Here we will report on the accuracy and approaches used in our recurrent genomic selection scheme using speed breeding.