OBJECTIVES: The authors conducted a theoretically guided study of access to inpatient or residential treatment among a probability sample of homeless adults with alcohol or drug use disorders in Houston, Texas. METHODS: The study used a cross-sectional, retrospective design with data collected from a multistage random sample of 797 homeless adults age 18 or older who were living in Houston shelters and streets in 1996. Structured, face-to-face interviews produced a screening diagnoses for alcohol and drug use disorders, treatment use data, and candidate predictors of treatment use. Logistic and linear regressional analyses were performed on the subset of 326 homeless persons with either alcohol or drug use disorders. RESULTS: 27.5% of persons with substance use disorder has accessed inpatient or residential treatment during the past year. Controlling for additional need factors such as comorbidity, persons having public health insurance and a history of treatment for substance abuse problems had greater odds of receiving at least one night of treatment. Contrary to expectation, contact with other service sectors was not predictive of treatment access. Schizophrenia and having a partner appeared to hinder access. Greater need for treatment was associated with fewer nights of treatment, suggesting retention difficulties. CONCLUSIONS: The study adds to previous findings on access to health care among homeless persons and highlights a pattern of disparities in substance abuse treatment access. Health insurance is important, but enhancing access to care involves more that economic considerations if homeless people are to receive the treatment they need. Referral relationships across different service sectors may require strengthening.