Background. This bioinformatics study was aimed at evaluating type 2 diabetes (T2D) and oral squamous cell carcinoma (OSCC) with regard to related immune cells and prognosis. Methods. We downloaded the data on OSCC from TCGA and for T2D from GEO database. Differentially expressed genes were analyzed, i.e., for OSCC genes with p   value < 0.01 , log 2 FC > 0 ; and for T2D, genes with p   value < 0.05 , log 2 FC > 0 . The intersected genes between OSCC and T2D were cross-talk genes. The expression values of immune-related genes in case samples in OSCC and T2D were assessed and underwent multivariate and univariate analysis (Cox-PH model). The intersection between the immune genes and cross-talk genes was taken and further analyzed by recursive feature elimination (RFE), survival analysis, and ROC analysis. Results. 1008 cross-talk genes were acquired, including 28 common upregulated, 440 common downregulated, and 540 differently regulated DEGs. We extracted the gene expression value of 782 immune-related genes, of which seven increased immune cells were obtained. From the results, plasmacytoid dendritic cells and effector memory CD8 T cells were highly negatively correlated in both OSCC and T2D. After estimating a low- and high-risk model for survival, we found that activated dendritic cell was significantly different between high and low groups ( p = 0.0095 ), followed by plasmacytoid dendritic cell. We integrated DE_Immune genes set 1 and DE_Immune genes set 2 and eight key immune-related cross-talk genes (C1QC, ABCD1, NOS2, PDIA4, IL1RN, ALOX15, CSE1L, and PSMC4) were evaluated. After ROC analysis, we obtained that ABCD1, C1QC, CSE1L, and PSMC4 had higher classification and prediction effects on OSCC and T2D. Conclusion. This study revealed a close relationship between T2D and OSCC. Thereby, plasmacytoid dendritic cell and activated dendritic cell-related genes were associated with the survival of T2D-related OSCC, while ABCD1, C1QC, CSE1L, and PSMC4 were the most important immune-related cross-talk genes.