Polycystic Ovary Syndrome (PCOS) is endocrine disorder effecting many women’s their in reproductive age system. This may result infertility and an ovulation, and it also causes hormonal imbalance which leads to a delayed or even absent menstrual cycle. Approximately 5-10 % of reproductive age (15-49 years) women are suffered from this problem. Women who are being suffered from this disease are being effected from following symptoms weight gain, facial hair growth, acne, hair loss, skin darkening and irregular periods. Now a days we have so many methodologies and treatments to predict other diseases in earlier stage, but when it comes to PCOS the existing treatments are insufficient to predict this disease in earlier stage. PCOS due to overlapping the follicles, inheritance of the equipment lack of operator knowledge as it is large experiment dependent procedure To deal with this problem infertility, diabetes mellitus, cardiovascular diseases, our proposed system which can help early detection and prediction of PCOS treatment from an optimal and minimal set of parameters. In this paper we are going to work with some machine learning algorithms those are Ad boost, Cat boost, XGboost, Gboost and Bagging. All these techniques are tested with applying Recursive Feature Elimination (RFE) methodology.