Ovarian cancer is a highly heterogeneous gynecological malignancy that seriously affects the survival and prognosis of female patients. Single-cell sequencing and transcriptome analysis can effectively characterize tumor heterogeneity to better study the mechanism of occurrence and development. In this study, we identified differentially expressed genes with different differentiation outcomes of tumor cells by analyzing a single-cell dataset. Based on the differentially expressed genes, we explored the differences in function and tumor microenvironment among clusters via consensus clustering. Meanwhile, WGCNA was employed to obtain key genes related to ovarian cancer. On the basis of the TCGA and GEO datasets, we constructed a risk model consisting of 7 genes using the LASSO regression model, and successfully verified that the model was characterized as an independent prognostic factor, efficiently predicting the survival prognosis of patients. In addition, immune signature analysis showed that patients in the high-risk group exhibited lower anti-tumor immune cell infiltration and immunosuppressive status, and had poorer responsiveness to chemotherapeutic drugs and immunotherapy. In conclusion, our study provided a 7-gene prognostic model based on the heterogeneity of OC cells for ovarian cancer patients, which could effectively predict the prognosis of patients and identify the immune microenvironment status of patients.