Objective: The interaction between tumor immune microenvironment (TIME) and malignant tumor cells plays a crucial role in the occurrence and development of tumors. This study aimed to establish and validate a prognostic model based on TIME characteristics for prognosis prediction and personalized treatment guidance of patients with endometrial cancer (EC). Methods: 67 EC patients who underwent surgery and TIME detection between January 2018 and December 2022 at Peking University People’s Hospital were included. A prognostic model was established according to the density of CD3+ cell and CD8+ cell. 200 EC patients as the validation set, we used immunohistochemical(IHC) markers CD3+cell and CD8+ cell, verified the accuracy of the prognostic model. Results: (1) Time was detected by multiplex immunofluorescence(mIF) in 67 EC patients from Peking University People’s Hospital, There were statistical difference between the Recurrence group and Non-Recurrence group in density of PD-L1+ cell, CD8+ cell, CD68+CD163+ cell, CD3+ cell and CD56+ cell. Among them, the difference of CD3+ cell was the most significant( P=0.004); (2)In 514 EC patients from TCGA database, There was statistical difference between the Recurrence group and Non-Recurrence group in CD8+ cell, regulatory T cells (Tregs) and Dendritic cells activated(DC activated)，Among them, the difference of CD8+ cell was the most significant( P=0.000); (3) CD3+cell and CD8+cell were selected as modeling factors to establish a prognostic model, the patients were divided into Cluster 1 (n=17), Cluster 2 (n=39) and Cluster 3 (n=11). Survival analysis difference( P=0.006) of the three groups were statistically significant; (4)The three Clusters has differences in other immune cells, molecular typing and clinicopathological characteristics; (5) 200 EC patients from Peking University People’s Hospital as the validation set, using IHC markers CD3+cell and CD8+ cell, the patients were divided into Cluster 1, Cluster 2, and Cluster 3 for survival analysis, which had statistical significance ( P=0.000) and verified the accuracy of the prognostic model. Conclusion: Based on TIME, this study found two immune cells that are significantly related to the prognosis of EC, and established an EC prognostic model based on this, with good prediction efficiency and accuracy.