Background: Colon adenocarcinoma (COAD) is one of the most common malignant tumors, with high incidence and mortality rates worldwide. Reliable prognostic biomarkers are needed to guide clinical practice. Methods: Comprehensive gene expression with alternative splicing (AS) profiles for each patient were downloaded using the SpliceSeq database from The Cancer Genome Atlas. Cox regression analysis was conducted to screen for prognostic AS events. The R package limma was used to screen differentially expressed genes (DEGs) between normal and tumor samples in the COAD cohort. A Venn plot analysis was performed between DEGs and prognostic AS events, and the DEGs that co-occurred with prognostic AS events (DEGAS) were identified. The top 30 most-connected DEGAS in proteinâprotein interaction analysis were identified through Cox proportional hazards regression to establish prognostic models. Results: In total, 350 patients were included in the study. A total of 22,451 AS events were detected, of which 2,004 from 1,439 genes were significantly associated with survival time. By overlapping these 1,439 genes with 6,455 DEGs, 211 DEGs with AS events were identified. After construction of the proteinâprotein interaction network, the top 30 hub genes were included in a multivariate analysis. Finally, a risk score based on 12 genes associated with overall survival was established (P < 0.05). The area under the curve was 0.782. The risk score was an independent predictor (P < 0.001). Conclusions: By exploring survival-associated AS events, a powerful prognostic predictor consisting of 12 DEGAS was built. This study aims to propose a novel method to provide treatment targets for COAD and guide clinical practice in the future.