Polysubstance use (i.e., simultaneous or sequential use of different psychoactive substances) is associated with increases in the risk of severe health problems and social impairments. The present study leverages community-representative, long-term longitudinal data from an urban cohort to assess: (a) the prevalence and continuation of polysubstance use between adolescence and early adulthood; (b) different patterns of polysubstance use (i.e., combinations of substances) in early adulthood; and (c) childhood risk factors for polysubstance use in early adulthood. At age 20 (n = 1,180), respondents provided comprehensive self-reported information on past-year substance use, including use of legal and illicit substances (e.g., cannabinoids, stimulants, and hallucinogens), and nonmedical use of prescription drugs (e.g., opioids, tranquilizers). In adolescence (ages 13-17), limited versions of this questionnaire were administered. In childhood (ages 7-11), potential risk factors, including individual-level factors (e.g., sensation-seeking, low self-control, aggression, and internalizing symptoms) and social-environmental factors (e.g., social stressors, exposure to others' substance use), were assessed. We fitted latent class models to identify classes of participants with different substance use profiles in early adulthood. The results show that polysubstance use increased between early adolescence and early adulthood. The continuation of polysubstance use was common (stability between all adjacent assessments: odds ratio >7). At age 20, more than one-third of participants reported polysubstance use (involving illicit substances, nonmedical use of prescription drugs, and cannabidiol). Four latent classes with polysubstance use were identified: (1) broad spectrum of substances; (2) cannabis and club drugs; (3) cannabis and the nonmedical use of prescription drugs; and (4) different cannabinoids. Risk factors for any polysubstance use included childhood sensation-seeking and exposure to others' substance use; some childhood risk factors were differentially associated with the four classes (e.g., low self-control in childhood was associated with an increased likelihood of being in the broad spectrum class). The classes also differed with regard to socio-demographic factors. This study revealed that polysubstance use is a widespread and multifaceted phenomenon that typically emerges during adolescence. To facilitate the design of tailored prevention mechanisms, the heterogeneity of polysubstance use and respective socio-demographic and developmental precursors need to be considered.