Iran is facing unprecedented dual drug use and HIV epidemics. We conducted a cross-sectional survey to obtain HIV prevalence and risk behavior data from injection drug users (IDU) in Tehran. We used respondent-driven sampling (RDS) to recruit IDU through successive waves starting with 24 "seeds," conducted anonymous face-to-face interviews and HIV testing and counseling, and used RDSAT to adjust data. During 44 weeks, 1,726 study referral coupons resulted in 645 (37 %) IDU referrals, of whom 548 (85 %) were enrolled. From those enrolled, 84 % were incarcerated, 47 % employed, 55 % single, 27 % under 30 years of age, and 26 % homeless. The adjusted HIV prevalence was 26.6 % (95 % confidence interval 21.3-32.1), and was higher among certain IDU subgroups (e.g., those who sharing injection paraphernalia). Our estimates of HIV prevalence were higher than some other estimates; however, repeated surveys using similar methodology are needed to monitor the trend of HIV epidemic over time.