Adolescent girls in the United States and around the world are at a heightened risk for sexually transmitted diseases (STDs), including human immunodeficiency virus (HIV). To determine the efficacy of a skill-based HIV/STD risk-reduction intervention in reducing self-reported unprotected sexual intercourse among African American and Latino adolescent girls. Randomized controlled trial with 3-, 6-, and 12-month follow-ups. Sexually experienced African American and Latino adolescent girls recruited from the adolescent medicine clinic of a childrens hospital serving a low-income inner-city community (N = 682, mean age, 15.5 years); 88.6% were retained at the 12-month follow-up. Three 250-minute interventions based on cognitive-behavioral theories and elicitation research: an information-based HIV/STD intervention provided information necessary to practice safer sex; a skill-based HIV/STD intervention provided information and taught skills necessary to practice safer sex; or a health-promotion control intervention concerned with health issues unrelated to sexual behavior. Primary outcome measure was self-reported frequency of unprotected sexual intercourse; secondary outcomes included the frequency of sexual intercourse while intoxicated, the number of sexual partners, biologically confirmed STDs, and theoretical mediator variables, including the intention to use condoms, beliefs about using condoms, and condom-use knowledge. No differences between the information intervention and the health control intervention were statistically significant. Skills-intervention participants (mean [SE], 2.27 [0.81]) reported less unprotected sexual intercourse at the 12-month follow-up than did information-intervention participants (mean [SE], 4.04 [0.80]; P = .03), or health control-intervention participants (mean [SE], 5.05 [0.81]; P = .002). At the 12-month follow-up, skills-intervention participants (mean [SE], 0.91 [0.05]) reported fewer sexual partners (P = .04) compared with health control-intervention participants (mean [SE], 1.04 [0.05]) and were less likely to test positive for STD (mean [SE], 10.5% [2.9%]) than were health control-intervention participants (mean [SE], 18.2% [2.8%]; P = .05). No differences in the frequency of unprotected sexual intercourse, the number of partners, or the rate of STD were observed at the 3- or 6-month follow-up between skill-intervention participants and information-intervention or health control-intervention participants. Skill-based HIV/STD interventions can reduce sexual risk behaviors and STD rate among African American and Latino adolescent girls in clinic settings.