There has been a long-standing debate about whether cheating is a stable behavior across various situations. However, there is a notable gap in our understanding about whether children's cheating behavior could exhibit stability over time. Moreover, research on the cognitive correlates of children's cheating is limited, yet exploring these cognitive factors is essential for understanding how children make (dis)honest decisions. This study aimed to test the longitudinal stability in children's cheating tendency and frequency and to explore the cognitive underpinnings of cheating behavior (theory of mind, inhibitory control, and free will belief). The study involved 100 children aged 3 to 6 years who were initially tested at Time 1, and 89 of these participants were retested at Time 2 approximately 1 year later. Cheating behavior was measured using a die-rolling game over Zoom, and three different cognitive abilities were measured. The results indicated that children's cheating tendency was stable over a year-long interval, whereas cheating frequency did not show longitudinal stability. Moreover, the study found that free will belief was related to cheating behavior, whereas theory of mind and inhibitory control were not. Specifically, children's belief in the free will to inhibit their desires, rather than their actual ability to inhibit the desires, was associated with a reduced frequency and likelihood of cheating. The findings can provide insight into the developmental origin of children's decisions to refrain from cheating.