Despite life-threatening consequences of pediatric cancer, treatment non-adherence rates range from 10-75%. Given the neurocognitive effects of chemotherapy it is important to study medication adherence to prevent disease relapse. The current study examined potential predictors of medication adherence among pediatric cancer patients. Survey data was collected from 104 parents of pediatric cancer patients through online and social media portals, family support groups, pediatric cancer websites, and foundations. Informed by the McMaster Model of Family Functioning, it was hypothesized that poor medication adherence would be predicted by multiple factors using multiple regression analysis. Results indicated poor communication (p = 0.049), behavioral control (p < 0.001), and affective responsiveness (p = 0.005) are predictors of poor medication adherence. Poor problem solving (p = 0.016) and higher degree of effort (p < 0.001) predict higher medication adherence. Post-hoc analysis indicated that when controlling for problem solving, child's sex had a significant effect on medication adherence. There was a significant effect of time since diagnosis on medication adherence after controlling for the effect of problem solving, communication, behavior control, and condition management effort. There was a significant effect of family's income on medication adherence after controlling for the effect of problem solving, communication, affective responsiveness, and condition management effort. In the current study, several predictors of adherence were identified. This knowledge helps medical and psychological professionals to work collaboratively to promote adherence. Future research is needed to further understand multicultural factors that may predict poor adherence and how specific pediatric cancer diagnoses impact family functioning.