Objective To investigate the value of ultrasound and MRI in the diagnosis and differentiation of benign and malignant ovarian tumors. Methods Eighty patients with ovarian cancer confirmed by operation and pathology were studied. All patients underwent ultrasound and MRI examination before operation. The diagnostic results were compared with the results of operation and pathology. The accuracy rates of ultrasound and MRI in differentiating benign and malignant ovarian tumors were analyzed, and the imaging findings of the two were also analyzed. Results The accuracy rate of preoperative qualitative diagnosis, and accuracy rates in diagnosis of benign and malignant ovarian tumors of MRI (90.00%, 96.08%, 93.55%) were slightly higher than those of ultrasound (P>0.05). Ultrasonic images showed that the boundaries of benign tumors were clear, the internal blood vessels were sparse, the branches were simple, the run was regular, blood vessels were smooth and the number was not significantly increased; the boundaries of malignant tumors were blurred, intratumoral blood vessels were rich, branches were in disorder, the run was irregularand was in stars form or cluster form, the number was increased significantly, and tortuosity was obvious. MRI image showed that benign tumors were solitary, the majority were cystic, cystic wall was thin and homogenous, most signals were homogenously enhanced, only part of the adjacent tissues were invaded, there was low signal on T1WI and high signal on T2WI, but enhanced scan showed no obvious enhancement; most malignant tumors were located in bilateral ovarian, mainly were cystic and solitary, cystic wall is thick, the intracystic compartment was inhomogenous, with pelvic wall, peritoneum, ascites and pelvic lymphadenectasis, there was equal-low signal on T1WI and mixed slightly high signal on T2WI, annular or uneven enhancement. Conclusion The ultrasound and MRI findings of ovarian tumors are obvious, which can complement each other in screening and differentiation of benign and malignant ovarian tumors, and can be used to improve the accuracy of clinical diagnosis.