Objective: This study aimed to determine whether on ultra-long range (ULR) flights, perceived workload is an independent predictor of cabin crew fatigue at top-of-descent (TOD) and if so, to what degree it is associated with cabin crew fatigue relative to sleep-related factors. Background: Current ULR scheduling for cabin crew is predominantly based on flight crew data. However, cabin crew workload is very different in nature to that of flight crew. Method: Fifty-five cabin crew wore an actigraph and completed a sleep/duty diary to monitor sleep during a ULR trip. At TOD, crewmembers completed a 5-min Psychomotor Vigilance Task (PVT), rated their sleepiness (Karolinska Sleepiness Scale) and fatigue (Samn-Perelli Crew Status Check), and after landing their workload (NASA Task Load Index). Results: When workload was perceived as higher, crewmembers felt more sleepy and fatigued and had more PVT lapses at TOD. The effect of workload on sleepiness was larger (Cohen's f(2) = .27) than the duration of wakefulness (Cohen's f(2) = .14), but the effect of workload on fatigue (Cohen's f(2) = .17) was smaller than the duration of wakefulness (Cohen's f(2) = .24). Lapses were not associated with sleep history, whereas workload had a small effect (Cohen's f(2) = .14). Conclusion: Workload as a fatigue factor for cabin crew warrants ongoing monitoring. This can be achieved by including a workload question in fatigue reports as an essential component in Fatigue Risk Management Systems.