Fusarium wilt (FW) caused by Fusarium oxysporum f. sp. elaeidis is the main disease of oil palm in Africa and is a latent threat to other producing regions. A long-term work based on a genetic survey of FW resistance using field observations and pre-nursery screening tests with inoculation resulted in FW resistant planting material whose resistance has not been overcome in 40 years. The genetic determinism of FW resistance has never been studied. Such information would eventually enable marker assisted selection (MAS) to ensure continuous improvement of planting material, through multi-trait selection (yield and disease or multiple diseases) and diversification of the genetic base. In this study, we used a pedigree-based QTL mapping approach that took advantage of the extensive pre-nursery FW resistance data recorded in the framework of a reciprocal recurrent selection program in Benin Republic. Eight QTL regions were mapped in two major genetic backgrounds and favorable QTL alleles were identified in the current breeding populations. The QTL pattern was specific to the population studied in terms of number, location and genetic variance explained, highlighting the different genetic architecture of FW resistance depending on the genetic background. To investigate a putative trade-off between FW resistance and yield, FW resistance QTL genotypes were predicted in a population that was formerly evaluated for yield in the field. The effects on yield of a FW QTL were offset in the commercial planting material thanks to different effects on different yield components in both heterotic groups. These results shed light on FW resistance in oil palm and provide valuable information for the implementation of MAS in the breeding program. Considering that prenursery screening tests are widely used, the approach presented here could be implemented in other breeding programs to provide further insights into FW resistance, especially in the context of wider genetic diversity.