The present study aimed to identify the recurrence‑associated genes in colon cancer, which may provide theoretical evidence for the development of novel methods to prevent tumor recurrence. Colon cancer and matched normal samples microarray data (E‑GEOD‑39582) were downloaded from ArrayExpress. Genes with significant variation were identified, followed by the screening of differentially expressed genes (DEGs). Subsequently, the co‑expression network of DEGs was constructed using the weighted correlation network analysis (WGCNA) method, which was verified using the validation dataset. The significant modules associated with recurrence in the network were subsequently screened and verified in another independent dataset E‑GEOD‑33113. Function and pathway enrichment analyses were also conducted to determine the roles of selected genes. Survival analysis was performed to identify the association between these genes and survival. A total of 434 DEGs were identified in the colon samples, and stress‑associated endoplasmic reticulum protein family member 2 (SERP2) and long non‑coding RNA‑0219 (LINC0219) were determined to be the vital DEGs between all the three sub‑type groups with different clinical features. The brown module was identified to be the most significant module in the co‑expression network associated with the recurrence of colon cancer, which was verified in the E‑GEOD‑33113 dataset. Top 10 genes in the brown module, including EGF containing fibulin like extracellular matrix protein 2 (EFEMP2), fibrillin 1 (FBN1) and secreted protein acidic and cysteine rich (SPARC) were also associated with survival time of colon cancer patients. Further analysis revealed that the function of cell adhesion, biological adhesion, extracellular matrix (ECM) organization, pathways of ECM‑receptor interaction and focal adhesion were the significantly changed terms in colon cancer. In conclusion, SERP2, EFEMP2, FBN1, SPARC, and LINC0219 were revealed to be the recurrence‑associated molecular and prognostic indicators in colon cancer by WGCNA co‑expression network analysis.