BACKGROUND The mortality experience of iron and steel workers from modern plants in developing countries has not been extensively described. METHODS Mortality at two Korean iron and steel manufacturing complexes was analyzed using Poisson regression methods with both direct and indirect standardization. Work histories were linked with a national mortality registry. Workers (44,974) hired beginning in 1968 were followed from 1992 to 2001. RESULTS The 806 deaths observed during 10 years of follow-up comprised 2% of the population at risk and represented a large healthy worker effect (HWE) for all causes (SMR = 0.59, 95% CI = 0.55-0.63) and for cancer (SMR = 0.79, 95% CI = 0.70-0.90). Mortality at subsidiaries was considerably higher than at the parent plants (SRR = 1.71, 95% CI = 1.47-1.99). Relative mortality rates declined with employment duration: > 20 years had significantly reduced mortality (SRR = 0.59, 95% CI = 0.43-0.82) compared to duration < 1 year (test for trend: P = 0.0006). Fatal injury deaths in the first year were highly elevated (SMR = 3.10, 95% CI = 2.17-4.26) declining to less than that expected after 5 years. Cancer mortality was elevated in stainless steel production (SRR = 3.26, 95% CI = 1.37-6.49) and overall mortality was elevated for work in plant maintenance departments (SRR = 1.17, 95% CI = 1.00-1.37), particularly for fatal injuries (SRR = 1.67, 95% CI = 1.29-2.14). All-cause mortality increased with employment duration in the steel-production departments, as did fatal injuries in material handling/construction. CONCLUSIONS This steelworker cohort exhibits excess mortality in some process areas. More detailed retrospective exposure assessment and future follow-up of this cohort will better define health risks in the modern iron and steel manufacturing.