Uveal melanoma (UM) is a highly malignant intraocular tumor. The imbalance of alternative splicing (AS) is a landmark of tumor initiation and progression. However, there are few studies of AS in UM. Thus, this study aimed to identify a new AS-based prognostic signature and reveal its relationship with tumor-infiltrating immune cells. Univariable Cox regression analysis identified survival-related AS events. The prognostic signature was constructed using the univariable and multivariable Cox regression analyses. Kaplan-Meier survival analysis, the proportional hazard model, and receiver operating characteristic curves verified its prognostic value. Single-sample gene set enrichment analysis was used to analyze immune cell enrichment. The correlation of the risk score with tumor-infiltrating immune cells and immune checkpoint blockade (ICB) genes was examined. We screened 2886 survival-related AS events, of which five were selected to build a prognostic predictor. The risk score was positively relevant with ICB key targets (HAVCR2, IDO1, and PDCD1) and the infiltration of T cells, MDSC, and activated B cells. We provided novel and effective indices, including a risk score and clinical nomogram, for prognostic prediction in UM and discussed the potential relationship between survival-related AS events and immune cell infiltration, which is crucial for developing immune-targeted therapy to improve prognosis.