This study used multiple levels of analysis to examine associations of alcohol and marijuana (AM) use with condom use, including the moderating role of partner type, in a general sample of 754 sexually active young adults (mean age = 18.6, SD = 0.63; 77% non-White). Multivariable models examined associations over the past 30 days (global) and at the most recent sexual event (event specific). Over the past 30 days, participants who had a recent casual partner and used both substances reported a higher proportion of condom-protected sexual events during this period compared to AM nonusers, whereas those who had a recent steady partner and used alcohol, marijuana, or both substances reported a lower proportion of condom-protected sexual events compared to AM nonusers. At the most recent sexual event, participants who used both substances were four times more likely than AM nonusers to have condom-protected sex if the event involved a casual partner; however, there were no significant associations if the event involved a steady partner. Marijuana use, particularly combined with alcohol use, deserves greater attention in the context of condom use. Future research should consider different levels of analysis and partner types to fully understand these complex associations.