Objective Because of the poor prognosis of colon adenocarcinoma (COAD), it is necessary to screen prognosis-related genes in COAD patients and establish a new prognostic risk assessment model. Methods COAD-related data from the cancer genome atlas (TCGA) and gene expression omnibus (GEO) were used as training and validation sets, respectively. Weighted gene co-expression network analysis (WGCNA), a Cox regression model and least absolute selection and shrinkage operator (LASSO) regression analysis were used to screen prognosis-related genes of COAD and establish a prognostic model. A receiver operating characteristic (ROC) curve was combined with a survival curve to verify the model accuracy, and a nomogram was constructed. Patients were divided into two groups by the median risk score. The immune cell proportion score (IPS) was used to evaluate the immunotherapy response of the two groups. Results A total of 15 feature genes were screened. The area under the ROC curve in the predictive model of COAD patients was >0. 6, and the survival rate of the high-risk group was significantly lower than that of the low-risk group (P< 0. 05), suggesting a good distinguishing ability for high- and low-risk COAD patients. Patients in the low-risk group had a higher IPS (P= 0. 026), indicating a better response to immunotherapy. Conclusions The model developed for COAD in this study has a good ability to predict the survival of patients at high and low risk of COAD.