Mitochondria are the main sites of oxidative metabolism and energy release of sugars, fats and amino acids in the body. According to studies, malignant tumor occurrence and development have been linked to abnormal mitochondrial energy metabolism (MEM). However, the feasible role of abnormal MEM in colon adenocarcinoma (COAD) is poorly understood. In this work, we obtained COAD patient data from The Cancer Genome Atlas (TCGA) as the training set, and GSE103479 from Gene Expression Omnibus (GEO) as the validation set. Combined with the mitochondrial energy metabolic pathway (MEMP)-related genes in Kyoto Encyclopedia of Genes and Genomes (KEGG) database, a risk prognostic model was constructed by utilizing Cox regression analysis to identify 6 feature genes (CYP4A11, PGM2, PKLR, PPARGC1A, CPT2 and ACAT2) that were significantly associated with MEMP in COAD. By stratifying the samples based on riskscore, two distinct groups, namely the high-and low-risk groups, were identified. The model demonstrated accurate assessment of the prognosis risk in COAD patients and exhibited independent prognostic capability, as evidenced by the survival curve and receiver operating characteristic (ROC) curve analysis. A nomogram was plotted based on clinical information and riskscore. We proved it could predict the survival time of COAD patients effectively combined with the calibration curve of risk prediction. Subsequently, based on the immune evaluation and mutation frequency analysis performed on COAD patients, patients in high-risk group had observably higher immune scores, immune activity and PDCD1 expression level than low-risk group. In general, the prognostic model developed using MEMP-related genes served as a valuable biomarker for forecasting the prognosis of COAD patients, which offered a reference for the prognosis evaluation and clinical cure of COAD patients.