The molecular classification of endometrial carcinoma (EC) has divided patients into four distinct prognostic subtypes based on the genomic abnormalities. However, in clinical application, the genomic copy number index lacks a simple and accurate detection method, and the method of using p53 immunohistochemistry or TP53 mutation as a surrogate for copy number status needs to be optimized. Here, we explored the application of genome instability (GI) index in molecular classification of EC. We assessed a retrospective cohort of EC from Beijing Cancer Hospital between 2011 and 2020. Genomic profiling of tumor tissues from EC patients was sequenced by OncoScreen Plus Panel of 520 cancer-related genes, and tumor immune microenvironment was evaluated using multiplex immunofluorescence staining. GI index was calculated as the sum of loss of heterozygosity, telomeric allelic imbalance, and large-scale state transitions scores. 143 patients were enrolled in this study, with the median follow-up time of 38.77 months. Molecular analysis firstly identified four different prognostic subgroups: POLE mutations (mut), MSI-H, TP53 wildtype, and TP53 mut. We also divided patients into two groups based on GI index: GI-high and GI-low, and patients with GI-high were found in three subgroups except for POLE mut. Further analysis showed that GI status had a more important effect on prognosis than TP53 mutation. Therefore, we proposed a three-group molecular classification based on POLE mutation and GI index: POLE mut (n=9, 6.5%), GI-low (n=116, 82.9%), and GI-high (n=15, 10.7%). Patients with GI-high had much poorer survival than those with GI-low [disease-free survival: hazard ratio (HR), 17.87, 95% CI, 6.67-50.00, P < 0.001; overall survival: HR, 16.75, 95% CI, 4.76-50.00, P < 0.001]. The application in TCGA-UCEC cohort containing 481 EC patients confirmed that this new method worked well for risk stratification of EC. Further analysis showed that patients in GI-high group had higher tumor grade, aneuploidy and lower T cell infiltrations. The one-stop method of GI index combined with POLE mutation offers more accurate and stable disease risk stratification and demonstrates the potential for guiding therapy for EC patients.