Abstract: Using Juglans mandshurica and Ulmus japonica as test materials, we examined the variability in nonstructural carbohydrates (NSC) concentrations in the branches with different basal diameters with a branch analysis method and explored potential errors in estimating the crownscale NSC content introduced from various sampling protocols. The results showed that organs significantly influenced the crown NSC concentrations for both species. The mean concentrations of the sum of soluble sugars and starch (TNC) of the leaves, new twigs, old branches, and dead branches were 17.6%, 12.6%, 5.7% and 2.9%, respectively. Most of the NSC concentrations in leaves and new twigs varied insignificantly with basal diameter, age, length and height of the branch. However, the NSC concentration in old branches increased significantly with decreasing the basal diameter, age and length of the branch, and with increasing the relative height of the branch. Among the branch traits, basal diameter was the best predictor for the NSC concentration of the old branch (the R² between 0.87 and 0.95). The mean TNC contents of leaves, new branches, and old branches for the two species accounted for 28%, 2% and 70% of the crown TNC content, respectively. Considering the effect of the spatial variability in the estimation of NSC content, we recommend the sampling protocol that applies the NSC concentration of new twigs and old branches with a diameter of 3 cm to up-scale the crown NSC content as a simple and practical method.