Ovarian cancer is the most lethal gynecologic malignancy which has high morbidity. Currently, histopathology, ultrasonic and CA125 detecting are the main diagnostic techniques for ovarian tissues. Though these methods have significantly increased the survival rate of patients with ovarian cancer, there is still a challenge in terms of distinguishing adenoma and early adenocarcinomas from benign hyperplastic polyps. So Raman spectroscopy was applied as a sensitive diagnostic alternative to identify pathologic changes (e. g., dysplasia) in ovarian tissue at the molecular level, using partial least-squares-discriminant analysis (PLS-DA) model. The subtle Raman variations among normal and cancerous ovarian tissues are associated with the transformation of cancerous tissues. Multivariate statistical method of partial least-squares-discriminant analysis (PLS-DA), together with the leave-one-patient-out cross-validation, is employed to build the discrimination model. In this research, we choose the corresponding LV (latent variables) numbers as 5, which has the lowest CV classification error. In this way, there is 3961% information of functional group captured. In addition, p-value is also calculated to compare and it is known that the first LV (p=107×10-13) has the most significant effect. Meanwhile, through the model we can know Raman spectroscopy associated with PLS-DA modeling provides highly specific signatures of various biomolecules, rendering a sensitivity of 862%, a specificity of 854%, and collectively a diagnostic accuracy of 852%. The results demonstrate that Raman spectroscopy can be used with PLS-DA model as a sensitive diagnostic alternative to identify pathologic changes in ovarian at the molecular level.