Background: The ability to diagnose or treat colorectal cancer (CRC) at an early stage could lead to the improved survival. The purpose of this study was to identify differentially expressed microRNAs (miRNAs) involved in the pathogenesis of CRC. These miRNAs may serve as the biomarkers or potential therapeutic targets for CRC early detection or treatment. Methods: The GSE39814 miRNA expression profile dataset was downloaded from the NCBI GEO database. MiRNAs were extracted from the exosome fractions of three normal fetal colon-derived cells and HCT116 and SW480 cell lines. Differentially expressed miRNAs between HCT116 and normal control cells and between the SW480 group and the normal controls were identified using the limma package and analyzed with the t-test. MiRNAs with vertical bar log(2) fold change (FC)vertical bar >= 2 and p < 0.05 were defined as differentially expressed. Cytoscape software was used to construct the integrated miRNA-miRNA-gene regulatory network and further identify the degree score of differentially expressed miRNAs. The putative target genes were predicted using the miRecords and MirWalk databases. Results: A total of 77 and 75 differentially expressed miRNAs were identified in HCT116 and SW480 cells, respectively. Three hundred sixty-nine and 301 miRNA target genes were screened from HCT116 and SW480 cells, and most of them (283 from HCT116 and 231 from SW480) are associated with pathogenesis of CRC by comparing with the Comparative Toxicogenomics Database (CTD) database. The integrated miRNA-miRNA-gene regulatory network identified the most significantly dysregulated miRNAs (degree >= 10). The potential target genes of the three highest degree miRNAs were predicted for analyzing their potency to support the development of diagnostic predictors or therapeutic strategies for CRC. Conclusions: Numerous miRNAs have perturbed expression in HCT116 and SW480 cell lines compared to controls. The assessment of three of the elevated miRNAs, miR-125a-3p/127-3p/320, with diagnostic or therapeutic potential could be clinically valuable for CRC.