BACKGROUND: The rise in melanoma-related mortality in Sweden has been less pronounced than the increase in incidence. Interventional activities aimed at early detection may have contributed to this discrepancy. METHODS: Individuals with malignant melanoma as the underlying cause of death between 1970 and 1996 (n = 7177) formed the basis of this study. Annual age-standardized mortality rates were calculated using the direct method of standardization with the Swedish population of 1970 as reference. Temporal trends in the standardized rates were evaluated using a log-linear model. The effects of age, period, and cohort on the mortality trends were estimated using a Poisson regression model. RESULTS: Since the mid-1980s, melanoma-related mortality in Sweden has leveled off, with no further increase during the last 10-15 years. The contribution to the mortality from noncutaneous melanoma was proportionally stable (20-25%) during the studied period. In females, a significant decrease in mortality from cutaneous melanoma was shown for the period of 1987-1996 with an estimated annual decrease of -2.3% (95% confidence interval: -4.3 to -0.3). This trend appeared to be more pronounced in the Stockholm-Gotland region. The observed trends were best explained with the age-period model in both genders. CONCLUSIONS: Melanoma-related mortality in Sweden has leveled off since the mid-1980s. During the period 1987-1996, a statistically significant downward trend was observed for females. This trend coincides with increased preventional activities.