Objective To study the infiltration pattern of bladder cancer immune cells and explore the relationship between immune cells and tumor prognosis. Methods The bladder cancer transcript data and corresponding clinical data were downloaded from The Cancer Genome Atlas (TCGA). CIBERSORT software deconvolution method was used to calculate the proportions of 22 kinds of immune cells. R software was used to analyze the immune cell infiltration pattern in each clinical stage. The relationship between each immune cell and prognosis was analyzed by Kaplan-Meier survival. Results A total of 433 cases of gene transcript data were downloaded from the TCGA database, including 414 cases of bladder cancer tissues and 19 normal tissues. After the data were corrected, the proportions of immune cells were calculated using CIBERSORT software. Screening was performed, and 146 cases of bladder cancer tissue and 4 cases of normal bladder tissue were obtained. Activated CD4+ memory T cells and CD8+ T cells had the lowest expression and M0 macrophages had the highest expression in clinical stage IV of bladder cancer. The bladder cancer patients with high expression of activated CD4+ memory T cells, CD8+ T cells and low expression of M0 macrophages had beneficial prognosis. Conclusion The bladder cancer patients with high expression of activated CD4+ memory T cells, CD8+ T cells and low expression of M0 macrophages might have better clinical prognosis.