Checkpoint inhibitors (CPIs) have been widely applied in the treatment of patients with bladder cancer (BLCA). However, there is still unmet need to dissect response predict biomarkers. To uncover CPI response-related marker genes in cancer cells, we utilized SCISSOR, integrating single-cell RNA and bulk RNA sequencing data. Transcriptomic and clinical data from IMvigor210, UNC-108, and BCAN/HCRN datasets were collected to evaluate and validate the identified biomarkers and signatures. Additionally, we analyzed TCGA-BLCA and local-BLCA RNA-seq data to investigate alternative splicing events (ASEs). Cell viability was assessed in T24 and UMUC3 cells with RBM17 upregulation or downregulation. Through SCISSOR analysis, we discovered that the expression levels of RBM17, TAP1, and PSMB8 were significantly associated with CPI response. Since PSMB8 displayed a highly positive correlation with TAP1, we developed a CPI response score (CRS) signature based on the expression profiles of RBM17 and TAP1. The CRS demonstrated robust predictive capacity in IMvigor210, UNC-108, and BCAN/HCRN datasets and was associated with higher tumor mutational burden (TMB), PD-L1 expression, and unique genomic features. Notably, RBM17 was not linked to the clinical outcomes of BLCA patients but positively correlated with BLCA cell proliferation in vitro. In the meantime, RBM17 was correlated with higher activity in core biological pathways, including antigen processing machinery, CD8 + T effector cells, cell cycle, DNA damage repair, epithelial-mesenchymal transition, histone regulation, and immune checkpoints. Moreover, the high-RBM17 group showed enrichment of LumU/Ba/sq subtypes but fewer FGFR3 alterations. Lastly, RBM17 significantly upregulated ASEs in BLCA samples, leading to higher neoantigen levels, a more inflamed tumor microenvironment, and improved CPI response. RBM17 is associated with higher ASEs and neoantigen levels, thereby potentiating the efficacy of CPI in BLCA. The established predictive signature, utilizing only two genes, has the potential to streamline clinical applications, providing a cost-effective alternative to expensive genomic, transcriptomic, and biological feature tests.