Background: Pathologic risk factors (PRF) are correlated with treatment and prognosis in cervical cancer patients treated with radical hysterectomy (RH). Machine learning methods can be used in medical research and practice as a decision-making tool. However, the method of pre-operative prediction to PRF is limited and the clinical availability of machine learning methods remains unknown in cervical cancer. Methods: The data set was collected from 1260 stage IB-IIA cervical cancer patients treated with RH. Tumor-related factors, and 75 blood parameters were included. The simulations involved 6 machine learning methods: Gradient Boosting Machine, RF (Random Forest), Cforest (Conditional Random Forest), etc. The prediction ability of the models was determined based on the area under receiver operating characteristic curve (AUC). Findings: The best results were obtained by the RF model in deep stromal infiltration prediction with an accuracy of 72.7% and the AUC of 0.792, and the Cforest model acquired an accuracy of 65.6% and the AUC of 0.675 in lymph vascular space invasion prediction. The accuracy of the model for lymph node metastasis prediction was less than 65%. Blood markers including D-dimer and uric acid were associated with PRF. Interpretation: Machine learning methods based on blood parameters can provide critical diagnostic prediction of postoperative PRF for cervical cancer, which may promote personalized treatment strategy. Fund: This work was supported by National Natural Science Foundation of China (81802619), CAMS Initiative for Innovative Medicine (No.2017-I2M-2-003), Special fund for China Cancer Foundation Beijing Hope Marathon (LC2018A13), and Graduate Education Reform Project of PUMC (10023201800302). Declaration of Interests: The authors have nothing to disclose. Ethics Approval Statement: The study was conducted in accordance with the ethical principles of the Declaration of CICAMS. This study was approved by the institutional review board of the center.