The detrimental effect of delays of retention on working memory has been a source of constant debate since the fifties. While studies in adults have proposed that temporal decay or interference underlie this effect, developmental change in the rate of forgetting is often mentioned to account for the developmental growth in working memory capacity. However, forgetting rate has never been examined in early childhood, and the few studies in older children led to divergent findings about the age-related change in the speed of forgetting. The present study documented for the first time this effect in early childhood. Using the shopping span task, children aged from 2;6 to 3;6 years observed the experimenter filling a bag with a series of plastic fruits and had to reproduce the series after various delays. Although their recall decreased with longer delays, the rate of forgetting on span was not affected by age, and the slope of linear regression between span and delay was similar across the three age groups. These findings suggest that forgetting rate did not evolve throughout early childhood, and as a consequence, the age-related increase in working memory capacity should be accounted for by changes in other processes than the rate of forgetting. Our results also pointed to important individual differences in early childhood that should be explored to broaden our understanding of working memory functioning in kindergarteners.