Introduction: Deep Learning is a rapidly growing technology that can find practical approaches to problems in various fields. Deep learning is assisting healthcare professionals and researchers to identify the hidden opportunities in data thereby serv ing the medical sector better. It also assists doctors to analyze any kind of disorders precisely and helps them to medicate the patients better, thus resulting in better medical decisions. Medical illness such as Polycystic Ovarian Syndrome (PCOS) does not have an effective diagnosis and proper treatment options. It is a prevalent endocrine disorder, which leads to the growth of ovarian cysts in child-bearing women, which further leads to infertility. Objective: To assist in the diagnosis of PCOS, deep learning methods such as Convolutional Neural Networks can be applied which produce effective results in image classification tasks. Methods: In this present study, an attempt has been made to compare the accuracies and other performance metrics of prior mentioned deep learning methods, and the problem of Over-fitting is discussed. Results and Conclusion: The main motto of using these deep learning methods is to precisely prognosticate whether a person is expected to have PCOS or not.