Children's screen time (ST) increased in recent years, but investigations of the content and context (e.g., parental presence, and device type) of ST in predicting early academic skills remains understudied. In this study of 127 four- and five-year-olds, we examined whether the educational content and contextual information about children's ST from time diary interviews at age 4 predicted direct assessments of academic skills at age five. Cluster analyses detected three unique groups: Cluster One was "lowest ST and highest non-educational TV, with little parental presence," Cluster Two was "moderate ST and highest educational mobile devices, with moderate parental presence," and Cluster Three was "highest total ST and moderate educational ST and high parental presence." Children in Cluster Two scored significantly higher in literacy skills at age five than peers in Cluster One and Cluster Three. Children in Cluster One scored significantly higher in spatial skills compared to those in Cluster Three. No significant associations were detected between children's ST and number skills. Findings suggest that educational content and contextual features of ST combine in meaningful ways and appear to better predict children's literacy and spatial skills compared to variable-centered approaches. Methodological implications of this research and future directions are discussed.