To investigate whether tobacco, alcohol, and their combined use are important risk factors for fire injuries, the authors analyzed data from a population-based case-control study in King County, Washington, between 1986 and 1987. Cases (n = 116) were households with at least one fatal or nonfatal unintentional residential fire injury reported to the Washington State Fire Incident Reporting System from 1984 through 1985. Controls were selected by random digit dialing (n = 256). After adjustment for household size, number of male household members, total household income, and education of the head of the household, the odds ratio for fire injury in households whose members collectively smoked 1-9 cigarettes per day was 1.5 (95% confidence interval (CI) 0.6-4.2) relative to households with no smokers; for 10-19 cigarettes per day, the odds ratio was 6.6 (95% CI 2.5-17.5), and for 20 or more cigarettes per day, it was 3.6 (95% CI 1.9-7.2). Although households with alcohol drinkers who consumed five or more drinks per occasion were found to be at increased risk in the crude analysis, multivariate analysis suggested that this was partly because drinkers tended to live in households with higher smoking levels. Thus, even though households with alcohol drinkers who consume five or more drinks per occasion may be at increased risk of residential fire injury, smoking appears to be the more important underlying risk factor.