There has been a growing interest in using the Internet to access a variety of populations for prevention education and health care. It is not clear that this is viable for homeless and other marginalized populations. The purpose of this study was to describe Internet access and use among a sample of homeless and indigent drug users. A brief questionnaire was developed to assess computer ownership, use, email, and Internet access and use. Participants were recruited through a community-based human immunodeficiency virus (HIV) prevention program. 265 participants completed the questionnaire. While 55% had ever used a computer in their lifetime, only 25% had ever owned a computer, 24% had ever had their own email account, 33% had access to a working computer in the last 30 days, 10% had sent or received email in the last 30 days, and 19% had surfed or accessed information on the Internet in the last 30 days. Logistic regression predicting lifetime use of a computer showed predictive factors of having completed high school (Odds Ratio [OR] = 4.5; 95% Confidence Interval [CI] = 2.41, 8.38) and number of days used amphetamines in the last 30 days, per 5 days of use (OR = 1.64; 95% CI = 1.18, 2.27). Inversely related were age, per 5 years of age (OR = 0.78; 95% CI = 0.65, 0.93) and number of days used marijuana in the last 30 days, per 5 days of use (OR = 0.97; 95% CI = 0.76, 0.99). Very low access and usage suggest that Internet-based public health education models be reexamined for this population. The association with amphetamine use may indicate a facilitating effect of amphetamine by heightening awareness and increasing wakefulness. (Authors)