Aim: Much information regarding predictors of illicit drug initiation and cessation is drawn from cross-sectional data. This paper aims to determine the longitudinal changes in factors associated with initiation and cessation of illicit drugs by young Australian women over a 3-year period. Participants: The sample was the cohort of young women moving from their mid- to late 20s, completing the Australian Longitudinal Study on Women's Health (ALSWH) survey in 2000 and 2003, who were either 'new' users or 'quitters' at the 2003 survey. Measurements: Crude and multivariate associations between changes in predictor variables and the probability of illicit drug initiation or cessation were evaluated. Variables significant in univariate analyses were used to create multivariable logistic regression models which predicted initiation and cessation of illicit drugs. Findings: All categories of smokers, except ex-smokers and those who adopted and quit smoking between surveys, were less likely to cease the use of illicit drugs. Women who became pregnant were more likely to cease illicit drug use. Women who continued to drink at levels described as long-/short-term risk and women who suffered continuing emotional abuse were less likely to cease use of illicit drugs. Conclusions: Longitudinal studies that examine factors associated with illicit drug initiation are best conducted in a cohort aged in their late teens to early 20s. Following the current cohort into their late 30s may further explain predictors of illicit drug cessation.