RNA sequencing has been extensively used to study specific gene expression patterns to discover potential key genes related to complex traits of interest in animals. Of note, a new regulatory mechanism builds a large-scale regulatory network among transcriptome, where lncRNAs act as competing endogenous RNAs (ceRNAs) to sponge miRNAs to regulate the expression of miRNA target genes post-transcriptionally. In this study, we sequenced the cDNA and sRNA libraries of nine liver samples from three Holstein cows during dry period, early lactation, and peak of lactation with HiSeq platform. As a result, we identified 665 genes, 57 miRNAs and 33 lncRNAs that displayed differential expression patterns across periods. Subsequently, a total of 41ceRNA pairs (lncRNA-mRNA) sharing 11 miRNAs were constructed including 30 differentially expressed genes. Importantly, 12 among them were presented in our large metabolic networks, and predicted to influence the lipid metabolism through insulin, PI3K-Akt, MAPK, AMPK, mTOR, and PPAR signaling pathways, thus, these genes were considered as the most promising candidates for milk fat formation. To our knowledge, this is first investigation to profile the ceRNA regulatory networks of liver transcriptome that could affect milk fat synthesis in bovine, providing a new view of the regulatory mechanism of RNAs.