Few studies have investigated the independent effects of domain-specific physical activity on mortality. We sought to investigate the association of physical activity performed in different domains of daily living on all-cause, cardiovascular (CVD) and cancer mortality. Using a prospective cohort design, 4,672 men and women, aged 25-74 years, who participated in the baseline examination of the MONICA/KORA Augsburg Survey 1989/1990 were classified according to their activity level (no, light, moderate, vigorous). Domains of self-reported physical activity (work, transportation, household, leisure time) and total activity were assessed by the validated MOSPA (MONICA Optional Study on Physical Activity) questionnaire. After a median follow-up of 17.8 years, a total of 995 deaths occurred, with 452 from CVD and 326 from cancer. For all-cause mortality, hazard ratios and 95% confidence interval (HR, 95% CI) of the highly active versus the inactive reference group were 0.69 (0.48-1.00) for work, 0.48 (0.36-0.65) for leisure time, and 0.73 (0.59-0.90) for total activity after multivariable adjustments. Reduced risks of CVD mortality were observed for high levels of work (0.54, 0.31-0.93), household (0.80, 0.54-1.19), leisure time (0.50, 0.31-0.79) and total activity (0.75, 0.55-1.03). Leisure time (0.36, 0.23-0.59) and total activity (0.62, 0.43-0.88) were associated with reduced risks of cancer mortality. Light household activity was related to lower all-cause (0.82, 0.71-0.95) and CVD (0.72, 0.58-0.89) mortality. No clear effects were found for transportation activities. Our findings suggest that work, household, leisure time and total physical activity, but not transportation activity, may protect from premature mortality.