This study tested components of the AIDS Risk Reduction Model (ARRM) for a sample of methamphetamine-using offenders in drug treatment. Analyses included the first two stages of the ARRM, problem recognition and intention to reduce risk (potential precursors to later possible behavior change), assessing predictors of intentions to increase condom use, reduce other sexual risk, and disinfect needles. Path analysis results showed potential applicability of the ARRM as a basis for intervention development for this population. There was a consistent effect of selfefficacy for risk reduction strategies, as well as direct or indirect effects of problem recognition factors (AIDS knowledge, peer norms), on the three intention indicators. Prior sex risk behavior (condom use) was directly negatively related to intention to use condoms; prior needle use was indirectly negatively related to intention to disinfect. Intention to use condoms was lower for females. Results can help identify areas for intervention development.