Summary Dissecting the genetic and genomic architecture of complex traits is essential to understand the forces maintaining the variation in phenotypic traits of ecological and economical importance. Whole‐genome resequencing data were used to generate high‐resolution polymorphic single nucleotide polymorphism (SNP) markers and genotype individuals from common gardens across the loblolly pine ( Pinus taeda ) natural range. Genome‐wide associations were tested with a large phenotypic dataset comprising 409 variables including morphological traits (height, diameter, carbon isotope discrimination, pitch canker resistance), and molecular traits such as metabolites and expression of xylem development genes. Our study identified 2335 new SNP × trait associations for the species, with many SNPs located in physical clusters in the genome of the species; and the genomic location of hotspots for metabolic × genotype associations. We found a highly polygenic basis of quantitative inheritance, with significant differences in number, effects size, genomic location and frequency of alleles contributing to variation in phenotypes in the different traits. While mutation‐selection balance might be shaping the genetic variation in metabolic traits, balancing selection is more likely to shape the variation in expression of xylem development genes. Our work contributes to the study of complex traits in nonmodel plant species by identifying associations at a whole‐genome level.