Prepubertal mammary development can be affected by nutrition partly through alterations in gene network expression. Quantitative PCR (qPCR) remains the most accurate method to measure mRNA expression but is subject to analytical errors that introduce variation. Thus, qPCR data normalization through the use of internal control genes (ICG) is required. The objective of this study was to mine microarray data (> 10,000 genes) from prepubertal mammary parenchyma and stroma to identify the most suitable ICG for normalization of qPCR. Tissue for RNA extraction was obtained from calves ( approximately 63 d old; n = 5/diet) fed a control (200 g/kg crude protein, 210 g/kg crude fat, fed at 441 g/d dry matter) or a high-protein milk replacer (280 g/kg crude protein, 200 g/kg crude fat, fed at 951 g/d dry matter). ICG were selected based on both absence of expression variation across treatment and of coregulation (gene network analysis). Genes evaluated were ubiquitously expressed transcript, protein phosphatase 1 regulatory (inhibitor) subunit 11 (PPP1R11), matrix metallopeptidase 14 (MMP14), ClpB caseinolytic peptidase B, SAPS domain family member 1 (SAPS1), mitochondrial GTPase 1 (MTG1), mitochondrial ribosomal protein L39, ribosomal protein S15a (RPS15A), and actin beta (ACTB). Network analysis demonstrated that MMP14 and ACTB are coregulated by v-myc myelocytomatosis viral oncogene, tumor protein p53, and potentially insulin-like growth factor 1. Pairwise comparison of expression ratios showed that ACTB, MMP14, and SAPS1 had the lowest stability and were unsuitable as ICG. PPP1R11, RPS15A, and MTG1 were the most stable among ICG tested. We conclude that the geometric mean of PPP1R11, RPS15A, and MTG1 is ideal for normalization of qPCR data in prepubertal bovine mammary tissue. This study provides a list of candidate ICG that could be used by researchers working in bovine mammary development and allied fields.