Background: Glioblastoma (GBM) and immunology are closely related, but its mechanism remains unclear. This study aimed to observe the causal inference between GBM and various immune cells by bidirectional Mendelian randomization (MR) analysis. Methods: We used immune cell and GBM data from the GWAS database. A total of 731 immunophenotypes, including four trait types and seven panels. For bidirectional MR analysis, Inverse Variance Weighted and False Discovery Rate (FDR) were both employed. Sensitivity analysis was also performed to make sure the results were reliable. Results: According to FDR, seven immunophenotypes associated with GBM risk: CD33br HLA DR+ AC (FDR = 0.009), CD38 on PB/PC (FDR = 0.046), CD66b on CD66b++ myeloid cell (FDR = 0.019), CD3 on CD39+ resting Treg (FDR = 0.009), HVEM on CM CD8br (FDR = 0.050), CD45 on CD33br HLA DR+ CD14dim (FDR = 0.027), and CD86 on CD62L+ myeloid DC (FDR = 0.048). In reverse MR analysis, GBM was found to be strongly associated with nine immunophenotypes based on FDR: BAFF-R on CD24+ CD27+ (FDR = 0.033), BAFF-R on IgD+ CD38- (FDR = 0.036), BAFF-R on IgD-CD38br (FDR = 0.039), BAFF-R on unsw mem (FDR = 0.048), BAFF-R on CD20- (FDR=0.012), HVEM on EM CD8br (FDR=0.036), CCR2 on myeloid DC (FDR = 0.035), CD45 on CD33-HLA DR+ (FDR = 0.004), and CD34 on HSC (FDR = 0.035). Conclusion: The current study confirmed the causal inference between 16 different immunophenotypes and GBM using genetic tools, providing an important foundation and guide for future immunological research and immunotherapy of GBM.