Silk of maize (Zea mays L.) contains diverse metabolites with complicated structures and functions, making it a great challenge to explore the mechanisms of metabolic regulation. Genome‐wide identification of silk‐preferential genes and investigation of their expression regulation provide an opportunity to reveal the regulatory networks of metabolism. Here, we applied the expression quantitative trait locus (eQTL) mapping on a maize natural population to explore the regulation of gene expression in unpollinated silk of maize. We obtained 3,985 silk‐preferential genes that were specifically or preferentially expressed in silk using our population. Silk‐preferential genes showed more obvious expression variations compared with broadly expressed genes that were ubiquitously expressed in most tissues. We found that trans‐eQTL regulation played a more important role for silk‐preferential genes compared to the broadly expressed genes. The relationship between 38 transcription factors and 85 target genes, including silk‐preferential genes, were detected. Finally, we constructed a transcriptional regulatory network around the silk‐preferential gene Bx10, which was proposed to be associated with response to abiotic stress and biotic stress. Taken together, this study deepened our understanding of transcriptome variation in maize silk and the expression regulation of silk‐preferential genes, enhancing the investigation of regulatory networks on metabolic pathways.