To explore the influence and effect of tumor microenvironment on the development of malignant mesothelioma using machine learning methods. 87 open cases were downloaded from the Cancer Genome Atlas database including transcriptome data, clinical data, and mutation data. The immune, stromal, and estimate scores were calculated for each case by using the ESTIMATE algorithm, and then the cases were grouped according to high and low stromal scores to predict all-cause survival in malignant mesothelioma cases. Their mutation data were analyzed to reveal the differences in mutated genes between the 2 groups, and then the von Willebrand factor (VWF) and FCRL3 genes were identified according to the intersection of DEGs and high-frequency mutated genes. Lastly, the correlation between VWF and the immune checkpoint of 22 kinds of immune cells was analyzed by using the CIBERSORT package of R software. A significant difference was found in the survival time of patients between the high and low stromal score groups. High expression of the VWF gene was negatively correlated with the prognosis of malignant mesothelioma, and the expression of VWF was positively correlated with naive B cells and activated CD4 memory T cells and negatively correlated with NK cells. The results revealed that high expression of VWF may involve in the development of malignant mesothelioma, and the anti-CTLA4 immune checkpoint treatment may have certain efficacy.