Esophageal squamous cell carcinoma (ESCC) is one of the leading causes of cancer-related mortality globally with a high risk of lymph node metastasis (LNM). In this study, weighted gene co-expression network analysis (WGCNA) showed the identification of 10 modules among which the significant module (turquoise), including 1352 genes, was correlated with LNM. A group 52 overlapping differentially expressed genes (DEGs) was identified based on the comparison of turquoise module with GSE23400 and GSE20347 datasets. Using Ctyohubba plugin, we identified 7 hub genes (ACTG2, SORBS1, MYH11, CXCL12, CNN1, IRS1 and CXCL8). IRS1 displayed significant correlation with metastasis. The decreased expression of IRS1 was also a predictor of poor OS of ESCC patients whereas the hub genes namely ACTG2, MYH11, CXCL8, CXCL12, IRS1 and CNN1 were associated with RFS of ESCC patients. These findings suggest that the altered expression of these hub genes are associated with prognosis and thus can be used as potential biomarkers for ESCC. Moreover, immunohistochemical staining and cell functional experiments displayed that the overexpression of IRS1 was negatively associated with metastasis in ESCC. In general, our research revealed several novel genes in ESCC especially the association of IRS1 with LNM in ESCC, which could provide novel insights into the initiation and progression of ESCC.