We established a mastitis model using exogenous infection of the mammary gland of Chinese Holstein cows with Staphylococcus aureus and extracted total RNA from S. aureus-infected and healthy mammary quarters. Differential expression of genes due to mastitis was evaluated using Affymetrix technology and results revealed a total of 1230 differentially expressed mRNAs. A subset of affected genes was verified via Q-PCR and pathway analysis. In addition, Solexa high-throughput sequencing technology was used to analyze profiles of miRNA in infected and healthy quarters. These analyses revealed a total of 52 differentially expressed miRNAs. A subset of those results was verified via Q-PCR. Bioinformatics techniques were used to predict and analyze the correlations among differentially expressed miRNA and mRNA. Results revealed a total of 329 pairs of negatively associated miRNA/mRNA, with 31 upregulated pairs of mRNA and 298 downregulated pairs of mRNA. Differential expression of miR-15a and interleukin-1 receptor-associated kinase-like 2 (IRAK2), were evaluated by western blot and luciferase reporter assays. We conclude that miR-15a and miR-15a target genes (IRAK2) constitute potential miRNA–mRNA regulatory pairs for use as biomarkers to predict a mastitis response.