Background HIV spread rapidly amongst injecting drug users (IDUs) in Bangkok in the late 1980s. In recent years, changes in the drugs injected by IDUs have been observed. We examined data from an HIV vaccine trial conducted amongst IDUs in Bangkok during 1999–2003 to describe drug injection practices, drugs injected, and determine if drug use choices altered the risk of incident HIV infection. Methods The AIDSVAX B/E HIV vaccine trial was a randomized, double-blind, placebo-controlled trial. At enrolment and every 6 months thereafter, HIV status and risk behaviour were assessed. A proportional hazards model was used to evaluate demographic characteristics, incarceration, drug injection practices, sexual activity, and drugs injected during follow-up as independent predictors of HIV infection. Results The proportion of participants injecting drugs, sharing needles, and injecting daily declined from baseline to month 36. Amongst participants who injected, the proportion injecting heroin declined (98.6–91.9%), whilst the proportions injecting methamphetamine (16.2–19.6%) and midazolam (9.9–31.9%) increased. HIV incidence was highest amongst participants injecting methamphetamine, 7.1 (95% CI, 5.4–9.2) per 100 person years. Injecting heroin and injecting methamphetamine were independently associated with incident HIV infection. Conclusions Amongst AIDSVAX B/E vaccine trial participants who injected drugs during follow-up, the proportion injecting heroin declined whilst the proportion injecting methamphetamine, midazolam, or combinations of these drugs increased. Controlling for heroin use and other risk factors, participants injecting methamphetamine were more likely to become HIV-infected than participants not injecting methamphetamine. Additional HIV prevention tools are urgently needed including tools that address methamphetamine use.