Given the established role of parental talk in the growth of math knowledge in preschoolers, there has been an increasing focus on identifying ways to promote parental math talk at this stage of child development. The current study investigated how parental math talk is affected by features of play materials and contexts. The features were manipulated along two dimensions: homogeneity (whether the toys were unique or included identical sets) and boundedness (whether the number of toys was restricted). Parent-child dyads (n = 75, child's age: 4-6) from China were randomly assigned to one of three experimental conditions: unique objects-unbounded range, homogeneous sets-unbounded range, and homogeneous sets-bounded range. In all conditions, dyads played games in two contexts that varied in the extent to which they were typically associated with math-party preparation and grocery shopping. As expected, more parental math talk was observed in the grocery shopping than in the party preparation context. Importantly, the manipulation of features within context also affected the amount and nature of parental math talk: homogeneity increased absolute magnitude talk, and boundedness increased relative magnitude talk. The results provide support for the cognitive alignment framework, underscoring the importance of aligning the features of materials with targeted concepts and demonstrating the possibility of impacting parental math talk through small manipulations of play materials.