Background Ovarian cancer is a highly fatal gynecological malignancy and new, more effective treatments are needed. Immunotherapy is gaining attention from researchers worldwide, although it has not proven to be consistently effective in the treatment of ovarian cancer. We studied the immune landscape of ovarian cancer patients to improve the efficacy of immunotherapy as a treatment option. Methods We obtained expression profiles, somatic mutation data, and clinical information from The Cancer Genome Atlas. Ovarian cancer was classified based on 29 immune-associated gene sets, which represented different immune cell types, functions, and pathways. Single-sample gene set enrichment (ssGSEA) was used to quantify the activity or enrichment levels of the gene sets in ovarian cancer, and the unsupervised machine learning method was used sort the classifications. Our classifications were validated using Gene Expression Omnibus datasets. Results We divided ovarian cancer into three subtypes according to the ssGSEA score: subtype 1 (low immunity), subtype 2 (median immunity), and subtype 3 (high immunity). Most tumor-infiltrating immune cells and immune checkpoint molecules were upgraded in subtype 3 compared with those in the other subtypes. The tumor mutation burden (TMB) was not significantly different among the three subtypes. However, patients with BRCA1 mutations were consistently detected in subtype 3. Furthermore, most immune signature pathways were hyperactivated in subtype 3, including T and B cell receptor signaling pathways, PD-L1 expression and PD-1 checkpoint pathway the NF-κB signaling pathway, Th17 cell differentiation and interleukin-17 signaling pathways, and the TNF signaling pathway. Conclusion Ovarian cancer subtypes that are based on immune biosignatures may contribute to the development of novel therapeutic treatment strategies for ovarian cancer.