Objective To investigate the antitumor effect of ubiquitin-specific protease 53（USP53） in colon adenocarcinoma（COAD） cells using bioinformatics methods combined with in vitro cellular experiments，since the clinical significance and potential mechanisms of USP53 in COAD remains unclear. Methods The expression data of USP53 from TCGA-COAD and GEO database （GSE39582 cohort） were analyzed，and KEGG，GO，and GSEA enrichment analyses were used to explore the biological processes and signaling pathways in which USP53 might be involved. The immune infiltration analysis and drug sensitivity analysis，along with the CLUE platform and molecular docking techniques，were used to predict the potential small-molecule compounds for treatment. A prognostic prediction model was established using 10 machine learning algorithms and 101 ensemble combinations，and in vitro experiments were performed to validate the function of USP53. Results USP53 exhibited an antitumor effect in COAD，and the low expression of USP53 was associated with poor prognosis（P<0.05）. The immune infiltration analysis showed that the low USP53 expression group had a higher level of immune infiltration，as well as higher TIDE and dysfunction scores（P<0.01）. In vitro experiments showed that USP53 overexpression inhibited the proliferation and migration of tumor cells and inhibited epithelial-mesenchymal transition （EMT）（P<0.05）. Conclusion USP53 exerts an antitumor effect by influencing the mechanisms such as EMT，apoptosis，cell cycle regulation，and tumor immune escape，and it may become a novel prognostic marker for COAD.