Researches focusing on the effects of alternative splicing (AS) on relapse of rectal cancer is little and signature based on the AS is blank. In this study, bioinformatic analysis was performed to identify and analyze the relapse-associated ASs, a signature was also constructed. In conclusion, 829 relapse-associated ASs of 676 mRNA were identified. 603 proteins with 2119 interactions were involved in the PPI (protein-protein interactions) network. 43 relapse-associated ASs and 64 SFs (splicing factors) with 160 interactions were indicated. Finally, we built a robust signature to predict the relapse of I–III rectal cancer with a high AUC (0.98) of ROC at 1 year. Based on the ASs involved in the signature, 4 molecular subgroups that could distinguish the relapse rate in diverse groups were identified. Our research provided an overview of relapse-associated ASs in I–III rectal cancer.