Objective: The aim of this prospective cohort study was to identify factors associated with fatigue in women with breast cancer who attended a 6-day rehabilitation course. Furthermore, to investigate level of fatigue 1 year after the rehabilitation course, after accounting for baseline fatigue levels. Methods: Between 2002 and 2005 we included 775 women with breast cancer and measured fatigue, emotional functioning and insomnia using the EORTC subscales and sociodemographics from a baseline questionnaire prior to participation in the rehabilitation course. Information on disease-specific and treatment factors was obtained from a nationwide clinical database. From the 12-month follow-up questionnaire we obtained level of fatigue. We estimated the influence of factors on the relative risk for a 10-points increase in fatigue score in generalised linear models. For the analysis of fatigue at 12 months follow-up we added fatigue level at baseline and relapse since baseline. Results: Higher tumour grade and insomnia predicted higher levels of fatigue, whereas better emotional functioning was associated with lower levels. After 12 months of follow-up fatigue levels were associated with level at baseline and with emotional functioning, insomnia, relapse and being on sick leave or unemployed. Emotional level, insomnia and high grade were associated with level of fatigue. After 1 year, a high level at baseline, insomnia and low emotional functioning was associated with greater fatigue. Conclusions: These longitudinal results support that cancer-related fatigue can be a persistent phenomenon and that focus on psychological distress and sleeping problems might be beneficial in these women.