Objective: To analyze the lympho-vascular space invasion (LVSI) in different molecular subtypes of the cancer genome atlas (TCGA) molecular subtypes of endometrial cancer (EC) and to evaluate the prognostic value of LVSI in EC patients with different molecular subtypes. Methods: A total of 258 patients diagnosed EC undergoing surgery in Peking University People's Hospital from January 2016 to June 2022 were analyzed retrospectively. Among 258 patients, 14 cases were classified as POLE-ultramutated subtype, 43 as high-microsatellite instability (MSI-H) subtype, 155 as copy-number low (CNL) subtype, and 46 as copy-number high (CNH) subtype. Fifty-four patients were positive for LVSI, while 203 tested negative. Results: (1) The incidence of LVSI was found to be highest in the CNH subtype (32.6%,15/46), followed by the MSI-H subtype (27.9%, 12/43), the CNL subtype (16.9%, 26/154), and the POLE-ultramutated subtype (1/14), with statistically significant differences (χ2=7.79, P=0.044). (2) Staging and deep myometrial invasion were higher in the LVSI positive group than those in the LVSI negative group (all P<0.05), except for the POLE-ultramutated subtype. The grade, lymph node metastasis, and the expression of nuclear antigen associated with cell proliferation (Ki-67) were significantly higher in LVSI positive patients than those in LVSI negative EC patients with both MSI-H and CNL subtypes (all P<0.05). In CNL subtypes patients, LVSI was also associated with age, histology subtype,and progesterone receptor (PR; all P<0.05). (3) Of the 257 EC patients, 25 cases recurred during the follow-up period, with a recurrence rate of 9.7% (25/257); among them, the recurrence rate of LVSI positive patients was 22.2% (12/54), which was significantly higher than those with LVSI negative (6.4%, 13/203; χ2=12.15, P<0.001). During the follow-up period, none of the 14 patients with POLE-ultramutated had recurrence; among CNL patients, the recurrence rate was 19.2% (5/26) in LVSI positive patients, which was significantly higher than that in LVSI negative ones (5.5%, 7/128; χ2=3.94, P=0.047); where as no difference were found in both MSI-H [recurrence rates in LVSI positive and negative patients were 2/12 and 9.7% (3/31), respectively] and CNH subtype [recurrence rates between LVSI positive and negative patients were 5/15 and 9.7% (3/31), respectively] EC patients (both P>0.05). After log-rank test, the 3-year recurrence free survival (RFS) rate were significantly lower in LVSI positive patients from CNL subtype and CNH subtype than those in LVSI negative patients (CNL: 80.8% vs 94.5%; CNH: 66.7% vs 90.3%; both P<0.05). (4) Lymph node metastasis (HR=6.93, 95%CI: 1.15-41.65; P=0.034) had a significant effect on the 3-year RFS rate of EC patients with MSI-H subtype. Multivariate analysis revealed that PR expression (HR=0.04, 95%CI: 0.01-0.14;P<0.001) was significantly associated with the 3-year RFS rate of CNL subtype patients. Conclusions: LVSI has the highest positivity rate in CNH subtype, followed by MSI-H subtype, CNL subtype, and the lowest positivity rate in POLE-ultramutated subtype. LVSI is significantly associated with poor prognosis in CNL subtype patients and may affect the prognosis of CNH subtype patients. However, LVSI is not an independent risk factor for recurrence across all four TCGA molecular subtypes.