To select genotypes with stable growth in diverse environments, tree breeders use multisite trials to evaluate genotypic stability and adaptation. Since genotype-by-environment (G × E) interaction effects vary with age, most multisite trials focus only on site effects, ignoring age effects. Liriodendron tulipifera trees are valuable due to their rapid growth and high-quality wood. Currently, multisite trials involving L. tulipifera plants are rare and the lack of data on G × E interaction effects impedes its selection. In this study, to explore the better performance and G × E pattern of L. tulipifera across ages, the growth traits (tree height, H; and diameter at breast height, DBH) of trees that were of five consecutive ages and grown on progeny-testing plantations were studied for 27 open-pollinated families at three sites. The results showed that the heritability of DBH was greater than that of H at almost all ages, and the individual breeding value ranking differed across sites and ages. The additive genetic correlations (r_A) between different site pairs were relatively small and varied with age, indicating an age trend for G × E, and showed a difference in traits. It was found that the absolute differences in some monthly average climatic indicators correlated with the G × E. Based on a comprehensive analysis considering stability and productivity, four elite families were identified. These results could aid in selecting stable, adaptable L. tulipifera genotypes and provide a reference for evaluating G × E interaction effects in multiage, multisite trials of other tree species.