This paper presents the results of a study evaluating the efficacy of a theory-based cognitive-behavioral intervention to reduce HIV risk among street-based crack and injection drug users not currently in drug treatment in Long Beach, California. A nine-session, 4-month enhanced intervention (including HIV counseling and testing) was compared to a two-session standard counseling and testing intervention developed by the National Institute on Drug Abuse (NIDA) in terms of their efficacy for reducing drug- and sex-related risk behaviors. The theory-based enhanced intervention rarely was found to be different from NIDA's standard counseling and testing intervention in reducing both drug- and sex-related risks, as indicated by cessation and/or reduction of drug use (measured by urine test and self-report), entry into drug treatment, and increased frequency of condom use. One of the few significant effects was that the enhanced intervention significantly increased injecting drug users' use of their own injection equipment. On the other hand, for both interventions, most risk behaviors were significantly reduced. It is concluded that the theory-based cognitive-behavioral intervention has limited advantage over the standard intervention in terms of both magnitude and frequency of HIV risk reduction achieved by high-risk, active drug users.