Senescent cells have been identified in the aging prostate, and the senescence-associated secretory phenotype might be linked to prostate cancer (PCa). Thus, we established a cellular senescence-related gene prognostic index (CSGPI) to predict metastasis and radioresistance in PCa.We used Lasso and Cox regression analysis to establish the CSGPI. Clinical correlation, external validation, functional enrichment analysis, drug and cell line analysis, and tumor immune environment analysis were conducted. All analyses were conducted with R version 3.6.3 and its suitable packages.We used ALCAM and ALDH2 to establish the CSGPI risk score. High-risk patients experienced a higher risk of metastasis than their counterparts (HR: 10.37, 95% CI 4.50-23.93, p < 0.001), consistent with the results in the TCGA database (HR: 1.60, 95% CI 1.03-2.47, p = 0.038). Furthermore, CSGPI had high diagnostic accuracy distinguishing radioresistance from no radioresistance (AUC: 0.938, 95% CI 0.834-1.000). GSEA showed that high-risk patients were highly associated with apoptosis, cell cycle, ribosome, base excision repair, aminoacyl-tRNA biosynthesis, and mismatch repair. For immune checkpoint analysis, we found that PDCD1LG2 and CD226 were expressed at significantly higher levels in patients with metastasis than in those without metastasis. In addition, higher expression of CD226 significantly increased the risk of metastasis (HR: 3.65, 95% CI 1.58-8.42, p = 0.006). We observed that AZD7762, PHA-793887, PI-103, and SNX-2112 might be sensitive to ALDH2 and ALCAM, and PC3 could be the potential cell line used to investigate the interaction among ALDH2, ALCAM, and the above drugs.We found that CSGPI might serve as an effective biomarker predicting metastasis probability and radioresistance for PCa and proposed that immune evasion was involved in the process of PCa metastasis.© 2022. The Author(s).