This dissertation addresses three novel aspects of children’s future thinking: First is a study of 3- to 5-year-olds’ capacity to think about the future across two different conceptual domains. Specifically, children had to think ahead to meet either a future physiological need (desire for food) or psychological need (avoiding boredom). Most future thinking tasks only require children to plan in one domain, this despite that future thinking is presumably domain general in humans. Children were better at addressing a future need for food than a future need for toys, with even 3-year-olds succeeding above chance. This study also served as an opportunity to replicate the results of a previous similar task (Atance et al., 2015) and improve the task by removing unnecessary components (social, pretense). Second is a study of 3- to 5-year-olds’ spontaneous talk (as a proxy for spontaneous thought) about the future and past within the context of a behavioural future thinking task. Spontaneous or involuntary thought about the past and future are ubiquitous in adult cognition. Few developmental studies have investigated past spontaneous thought, and none have investigated future spontaneous thought. Children of all ages spontaneously spoke about the future and past and some children even spontaneously solved the future thinking task. Further, children who spontaneously spoke more about the past and future were more likely to correctly answer an explicit test question. Third is an attempt at addressing a limitation in all existing behavioural future thinking tasks: That is, tasks used to assess the development of future thinking do not require children to think ahead about a future state of the world that differs from the present. Children could potentially be solving behavioural future thinking tasks without having to represent the future. However, representing a future that differs from the present is argued to be core to adaptive future thinking in humans. To overcome this limitation, we modified an existing task so that children could not succeed based on their representation of how the world currently is but, rather, how it will be at a future point in time. Four- to 7-year-olds all remembered the information required to solve the modified task; however, only 7-year-olds made a future-oriented decision more often than chance. With the task modification removed (so the correct answer for the present and future matched), even 4-year-olds succeeded above chance. These findings challenge the current accounts that suggest by age 4, children can reliably succeed in future thinking tasks. Taken together, this research program contributes new insights to the development of future thinking in early childhood and suggests directions for novel research.