The diverse set of skills that facilitate emotion regulation are crucial for adaptive developmental outcomes. The foundations for these skills are consolidated in social interactions during pre-school years. Dyadic pretend play is a social interaction that is posited to provide valuable opportunities for emotion regulation in this age group. However, our understanding of dyadic pretend play has not progressed to the point where this activity can be reliably structured to create a zone of proximal development for emotion regulation skills. In Study 1, a repeated-measures experiment was conducted to observe individual differences in emotion regulation during positively and negatively valenced dyadic pretend play. Thirty children between four and five years of age participated in two unstructured and two structured pretend play sessions. The structured sessions included a "pretend creature in a box" scenario where interaction with a pretend monster (negatively valenced condition), or kitten (positively valenced condition) was required in order for the pretend game to continue. It was expected that individual differences in the children's regulation of negative emotion tone and intensity would be observed in their response to the pretend monster. However, there were no significant differences in children's emotion tone and intensity, approach to the creature, and ability to continue the pretend play across the positively and negatively valenced conditions. Instead, individual differences were observed in the children's continuation of pretend play when information was gained about the pretend/reality status of the creature. The "pretend creature in a box" scenario was hypothesised to challenge children's regulation of anticipatory arousal elicited by uncertainty about the pretend/reality status of the creature. Study 2 examined this hypothesis in an experiment that explored emotion regulation across positively and negatively valenced pretend play, and conditions of high, low, and no uncertainty. Sixty children aged between four and five years participated in a structured pretend play game, which included an adaptation of the "pretend creature in a box" scenario. Anticipatory arousal elicited by uncertainty was manipulated by providing visual, verbal, and no information about the pretend/reality status of the creature. It was expected that individual differences in regulation of anticipatory arousal elicited by uncertainty would be observed in children's efforts to gain information about the pretend/reality status of the creature. Under conditions of uncertainty, a significantly greater proportion of children discontinued the pretend game by approaching the box and checking the pretend/reality status of the creature. Consistent with the results from Study 1, the intended emotional valence of the condition was not influential. Overall, this research programme highlights the challenges that occur in structuring dyadic pretend play to create a zone of proximal development for emotion regulation. Study 1 and Study 2 linked emotion regulation in pretend play with children's experience of anticipatory arousal elicited by uncertainty, and information seeking as an emotion regulation strategy. Study 2 demonstrated that pretend play has the potential to be utilised in the development of a set of skills that regulate the anticipatory arousal elicited by uncertainty. Variation in the particular strategies that a more experienced play partner might use to create and direct the experience of emotion and emotion regulation is a useful starting point for future research in this area.