Introgression from wild relatives is a potential source of beneficial diversity. The contribution of historic introgression to adaptive evolution and improvement of wheat following its origin and dissemination across the world remains unknown. Here, we used a reference wheat genome to generate a haplotype map including 7.3 million SNPs based on targeted resequencing of 890 diverse wheat landraces and cultivars, and tetraploid wild and domesticated relatives to identify genomic regions showing the signals of introgression from wild emmer. By analyzing SNP distribution relative to geography, historic environmental variables and improvement status (landraces vs. cultivars), we sought to assess the contribution of introgression to local adaptation and crop improvement, and to evaluate the effects of these factors on deleterious allele burden in wheat. The wild emmer source populations that contributed to gene flow differed across the wheat genome. It was lower in the A than in the B genome for northern wild emmer population, but it was higher in the A than the B genome for wild emmer from southern range of species distribution. On average, introgressed genomic regions (IGRs) comprised about 11.8% and 11.4% of genome per accession in landraces and cultivars. Genomic regions targeted by improvement selection and associated with environmental adaptation overlapped with 20.4% and 8% of detected IGRs, respectively, suggesting that wild emmer may have also contributed beneficial allelic diversity used during the development of locally adapted cultivars. By studying deleterious SNP distribution across the wheat genomes, we showed that improvement selection, environmental adaptation and introgression significantly reduced deleterious mutation burden in modern wheat. The overlap between the signals of gene flow, GWAS, improvement and environmental adaptation was also suggestive of adaptive introgression. Using the variance-component method we showed that alleles introduced from wild emmer explained substantial proportion of phenotypic variance for harvest weight (up to 30.9%), drought susceptibility (up to 22.5%) and plant height (up to 35%) traits. Our results provide evidence that historic gene flow from wild relative played an important role in shaping the agronomic phenotypes in modern wheat and likely broadened its adaptive potential. A detailed map of genome-wide introgression developed in our study can guide targeted deployment of wild relative diversity in wheat breeding programs. These efforts, besides introducing novel adaptive variation available for selection, hold great potential to reduce the deleterious mutation burden in the wheat genome and further accelerate breeding.