The authors study the relationship between disengagement from gangs and desistance from crime within a life-course criminological framework. Gang disengagement is conceptualized as the event of gang membership de-identification and the process of declining gang embeddedness. The authors examine the effects of both the event and the process of disengaging from gangs on (1) criminal desistance mechanisms and (2) criminal offending using longitudinal data and multilevel modeling. The authors found that disengaging from gangs is indirectly related to offending through less exposure to antisocial peers, less unstructured routine activities, less victimization, and more temperance. Gang disengagement is associated with decreased contemporaneous offending but does not predict future offending after controlling for desistance mechanisms. Evidence also suggests that those who leave gangs more quickly are less exposed to antisocial peers, and possess better work histories and psychosocial characteristics even while in the gang. The authors discuss implications for research on gangs and criminal desistance. Abstract published by arrangement with Taylor and Francis.