We investigated relationships between drinking, other drug use, and drug craving, using ecological momentary assessment (EMA), in a sample of polydrug users who were not heavy drinkers. In a prospective longitudinal cohort study, 114 heroin and cocaine users on methadone-maintenance treatment carried handheld electronic diaries during waking hours and were screened for drug and alcohol use for up to 25 weeks. Individuals who fulfilled the Diagnostic and Statistical Manual of Mental Disorders criteria for alcohol abuse or dependence were excluded. Participants responded to 2-5 random prompts per day to report on their moods, cravings, and activities and initiated entries when they used or acutely craved heroin or cocaine. Drinking alcohol was assessed in both types of entries. Breath alcohol was measured three times weekly. Participants reported drinking alcohol in 1.6% of random-prompt entries, 3.7% of event-contingent entries when craving cocaine and/or heroin, and 11.6% of event-contingent entries when using cocaine and/or heroin. Alcohol drinking was also associated with higher craving ratings and prestudy alcohol use. More drinking was detected by ambulatory self-report than by in-clinic breath testing. Even though we had screened out heavy drinkers from our sample of polydrug users, drinking was associated with heroin and cocaine craving and actual use.