Accurate and reliable diagnosis of cervical cancer is crucial for effective clinical treatment and prognosis evaluation. Given the low sensitivity of existing serum estrogen and tumor markers in assessing lesion features like parauterine invasion and lymph node metastasis, this study proposes a diagnostic method integrating multi-phase dynamic enhanced magnetic resonance imaging (DCE-MRI). Firstly, serum estradiol (E2), follicle-stimulating hormone (FSH), luteinizing hormone (LH), glycochain antigen 125 (CA125), glycochain antigen 19-9 (CA19-9) and DCE-MRI were detected to analyze the correlation between cervical cancer stage and serum levels of E2, FSH, LH, CA125 and CA19-9. Then, cervical biopsy results were used as the gold standard to analyze the diagnostic accuracy of DCE-MRI for cervical deep muscle invasion and lymph node metastasis. Finally, the serum levels of E2, FSH, LH, CA125, CA19-9, DCE-MRI diagnosis and combined diagnosis of cervical cancer were evaluated based on the receiver operating characteristic (ROC) curve. The experimental results showed that the AUC of combined diagnosis was 0.908, indicating that the combined diagnosis method effectively distinguished benign and malignant cervical lesions. Furthermore, DCE-MRI parameters K^trans and K^ep enriched the feature information, thereby improving the detection accuracy of lesions. This provides a novel approach for differentiating between benign and malignant cervical cancer and for determining its clinical stage.