Fusarium head blight (FHB) is a ravaging disease of bread and durum wheat production, and the accumulation of mycotoxins in grains due to this disease is imperiling food and feed safety and imposes risks to human and animal health. The development of resistance to FHB in bread wheat genotypes has been widely reported, in contrast to durum wheat which is still lagging behind with no resistant varieties available. Introgression of resistance genes from wild or cultivated relatives into durum wheat is one possible way to enlarge the durum resistance gene pool. Integrating molecular biology-based methods such as genome-wide association studies (GWAS) are beneficial for detecting the FHB resistance quantitative trait loci (QTL) present in various genetic sources in durum wheat. Therefore, we used GWAS to dissect the genetic control of FHB resistance in 603 multi-parental durum wheat breeding lines with resistance alleles derived from T. aestivum, T. dicoccum, and T. dicoccoides. We phenotyped the plant material for five years for FHB resistance, plant height, and anthesis date in artificially inoculated disease nurseries. Broad variation was found for all investigated traits, including many moderately FHB resistant experimental lines. Plant height influenced FHB resistance levels and led to co-localization of plant height and resistance QTL. Using plant height as a covariate in the GWAS model resulted in 12 QTLs associated with FHB resistance. For six QTLs, the durum cultivars contributed the resistance alleles, and the other six QTLs were derived from donor parents carrying QTL from resistant relatives. The strongest detected QTL is the height-independent resistance QTL Fhb1, derived from hexaploid wheat, successfully integrated into several durum genetic backgrounds. In addition to Fhb1, a QTL on chromosome 3A was identified, initially detected in the T. dicoccoides lines 'Mt. Hermon#22' (aka T. dicoccoides1A) and 'Mt. Gerizim#36' (aka T. dicoccoides52A).