Background: Sickness absence due to mental health problems (MHPs) is increasing in several European countries. However, little is known about return to work (RTW) for employees with MHPs. This prospective study aimed to identify predictors for RTW in employees sick-listed with MHPs. Methods: Employees were recruited when applying for sickness benefit due to MHPs from the Municipality of Copenhagen (n = 644). Information about age, gender, occupation, self-reported RTW expectancy, self-reported reason for absence and prior absence with MHPs was retrieved from application forms for sickness benefit. Each participant was followed-up in the National Register for Social Transfer Payments for a maximum period of 52 weeks to estimate time to RTW. Hazard ratios for RTW with 95% confidence intervals were calculated using Cox proportional regression analyses. Results: Employees sick-listed with self-reported stress/burnout returned to work faster than those with self-reported depression (HR = 0.76), and other MHPs (HR = 0.56). A positive RTW expectancy of the sick-listed person (HR = 1.27) and no prior absence with MHPs (HR = 1.29) were associated with a shorter time to RTW. Conclusion: Sickness absence due to self-reported stress/burnout, a positive RTW expectancy and no prior absence with MHPs predicted a shorter time to RTW among Danish employees sick listed with MHPs. Findings could help social insurance officers and other rehabilitation professionals to identify groups at high risk for prolonged absence.