The aim of the present study was to examine the predictive value of Return to Work Self-efficacy (RTWSE) on Return to Work (RTW) among employees undergoing chemotherapy for cancer and to examine the relative contribution of RTWSE as predictor variable compared to personal, health-related, illness-and treatment-related and work-related factors. Methods A sample of 114 sickness absent employees with various cancers (age 18-62) included in the study on average 33 days after initiating chemotherapy were followed for 15 months. Data sources included patient questionnaires (RTWSE, depression, fatigue, performance status), sociodemographic factors (age, sex, job type, and perceived support from the workplace), patient records (type of cancer, treatment intention, number of treatment modalities, time since diagnosis and time since initiation of chemotherapy), and Danish national registries (RTW and education). Associations between RTWSE at baseline and weeks until full RTW during 15-months follow-up were analyzed using Cox proportional hazards regression. Results In the univariate analysis, high RTWSE was associated with shorter time to RTW (Hazard Ratio (HR) 1.84, 95% confidence interval (CI) 1.12-3.03). In the multivariate model, RTWSE failed to reach statistical significance (HR 1.12, 95% CI 0.62-2.02), whereas female sex (HR 0.30, 95% CI 0.15-0.60) and receiving palliative treatment (HR 0.15, 95% CI 0.05-0.44) were significantly associated with later RTW. Conclusion Compared to other factors of significance, RTWSE was not the strongest predictor of RTW when examined among employees undergoing chemotherapy for cancer. Before using the RTWSE questionnaire to identify employees with cancer at risk of late RTW, it is important to recognize that the predictive value of RTWSE may be different for employees on sick leave due to cancer than for other sickness absence populations.