Globally, ovarian cancer is the 8th most familiar malignancy in females. Ovarian cancer growth is generally analyzed at a late phase of the sickness. Obtaining a clinical point of view about any medical complications is needed to getting ovarian malignant growth early as there is no viable screening test for this malignancy and most cases are merely found in their severe stages. Ovarian cancer is generally found at a serious stage, the endurance rate is distinctive among the various sorts of ovarian threat. A progressing form explanation empowered the usage of pathophysiological complications to investigate ovarian harmful development earlier. In this paper, we evaluated the accuracy, precision, and recall estimation of a proposed framework utilizing machine learning techniques to predict ovarian cancer types (Non-Cancer, Ovarian Cancer, Peritoneal Cancer, Fallopian Tube Cancer, and Ovarian LMP) from patients' pathophysiological complications. Our proposed machine learning classifiers - support vector machine, random forest, and XGBoost achieved 71%, 72%, and 69% accuracy respectively.