Abstract Aim The aim of the present study was to clarify prognostic role of angiogenesis in epithelial ovarian cancer. Methods Quantification of angiogenesis was performed by the Chalkley method after immunostaining of 175 epithelial ovarian cancer specimens with an antibody against CD34. Results The Chalkley count was categorised into two groups according to the median value: low <8 or high ⩾8. The low Chalkley count correlated significantly with serous and clear cell histological subtype of the tumour ( p < 0.0005), whereas there existed no association with FIGO (International Federation of Gynecology and Obstetrics) stage, histological grade, presence of primary residual tumour, age at diagnosis, or chemotherapy response. In univariate analysis, the high Chalkley count predicted poor overall survival in the subgroup of patients with FIGO stages III–IV tumours ( p = 0.007) but not in the entire study cohort. However, in multivariate analysis, the Chalkley count was found to be an independent predictor of death from ovarian cancer in the entire study cohort ( p = 0.044, RR = 1.50, 95% CI 1.01–2.21) as well as in the subgroup of FIGO stages III–IV tumours ( p = 0.046, RR = 1.58, 95% CI 1.01–2.46) together with the presence of primary residual tumour ( p < 0.0005, RR = 5.10, 95% CI 3.02–8.62, and p = 0.002, RR = 4.28, 95% CI 1.34–13.73, respectively). Conclusions The Chalkley count seems to be suitable for evaluation of angiogenesis and to have prognostic significance in ovarian cancer.