The objective of this study is to determine whether parent characteristics and practices differentiate between substance using and non-using children in Mexico. Data consisted of 52,171 children (i.e., 5th and 6th grade) who participated in the National Survey of Drug Use Among Students. The relative importance of student reported parenting factors in classifying lifetime substance use were assessed across outcomes with a logistic regression algorithm. Machine learning algorithms utilizing parent characteristics and practices classifed tobacco, marijuana, inhalants, and other substance use. Parenting factors were poor classifers of lifetime alcohol use. The algorithm generally selected important classifers broadly associated with parental illicit substance use and monitoring practices across child substance use outcomes. Findings suggest that leveraging parental infuences during childhood may be a high value point of intervention for preventing non-alcoholic substance use. Important classifers identifed by the current study represent potential targets for childhood substance use prevention efforts.