The present study examined the prospective relationship between parental education and adolescents' soft drink intake over 20 months, and possible mediating effects of adolescents' availability and accessibility of soft drinks at home. A total of 866 adolescents, with data on two time points in the Norwegian HEalth In Adolescents (HEIA) cohort study (2007–9), were included in the analyses. Data on intake and determinants of soft drinks were collected from adolescents and both parents by questionnaires. Mediation analyses using linear regression investigated the total and direct effects of parental education on adolescents' soft drink intake from the age of 11–13 years. In order to investigate prospective relationships, two models were set up to measure the (1) prediction and (2) change in consumption over 20 months. Possible mediation effects of availability and perceived accessibility at home were further examined in both models. The results showed that a lower level of parental education predicted a higher intake of soft drinks among adolescents after 20 months, and that higher perceived accessibility of soft drinks reported by adolescents and mothers explained 39 % of the total effect. No relationship was observed between parental education and the change in adolescents' intake of soft drinks over 20 months. Interventions aimed at families with low parental education should target the perceived accessibility of soft drinks at home in order to diminish social differences in adolescents' soft drink consumption.