Differences in children's mathematics knowledge are evident at kindergarten entry, favoring children who have greater access to economic resources. Fostering preschoolers' mathematics learning at home and in classroom settings, through games and other developmentally appropriate activities, is of great interest to educators, early childhood leaders, and policymakers. This cluster randomized trial examined the effects of a naturalistic, game-based mathematics intervention implemented in Head Start classrooms and examined whether including a family math component added value. A total of 573 children (64% Hispanic; 60% multilingual) were included from 66 classrooms which were randomly assigned to Classroom Math (CM), Classroom Math + Family Math (CM+FM), or business-as-usual (BAU). Results indicated that the family math component did add value to the classroom-based intervention as CM+FM resulted in a significant positive impact on children's mathematics knowledge relative to BAU, but CM did not. For preschoolers age 50+ months, both interventions had significant effects on children's mathematics knowledge relative to BAU, but CM+FM had a stronger effect (d = 0.36). The number of math games played was significantly associated with higher mathematics scores and the number of family math mini-books returned had a significant impact on children's spring scores, over and above the number of games played. The CM+FM intervention also had a significant effect on teachers' instructional practice (d = 0.79). Adding a family math component to a game-based classroom intervention resulted in positive impacts for preschoolers and seems to be an effective, ecologically valid intervention that fosters early mathematical competencies. [This paper was published in "Early Childhood Research Quarterly" v63 p43-58 2023.]