Forest productivity is a crucial integrator of ecosystem functions and services. Although the effects of landscape structure on species richness and stand structure have been extensively studied, how landscape structures affect forest productivity and their interactions with stand‐level attributes, especially in the context of considerable land use change, remain unclear. In this study, we investigated the effects of landscape structures (fragmentation, complexity, and heterogeneity) and their interactions with stand‐level attributes on forest productivity in the conterminous United States across three spatial scales (1–3 km), using an extensive forest inventory dataset from the national forest inventories (NFI) plots. Our results revealed that all landscape indices around selected forest plots significantly increased from 2006 to 2016. Across three scales, forest productivity and stand‐level attributes (number of trees, tree species richness, and structural diversity) exhibited unimodal relationships with landscape fragmentation and complexity, while generally showed positive correlations with landscape heterogeneity. The interactions between landscape structures and stand attributes enhanced the explanatory power of forest productivity. Landscape complexity directly or indirectly reduced forest productivity by decreasing the number of trees and tree species richness, whereas landscape fragmentation and heterogeneity had the opposite effects. Furthermore, landscape heterogeneity and stand age had relatively stronger total effects (the sum of direct and indirect effects) on forest productivity, and their strength increased modestly with spatial scales. However, tree species richness consistently had the lowest total effects. Our study elucidates the complex driving mechanisms of landscape patterns on forest productivity across spatial scales, providing a deeper understanding of ecosystem complexity and responses to accelerating land use changes.