Today, with the growing number of textual information sources, we have a lot of textual data.Researchers are faced with the problem of redundancy in accessing useful information and have to search through large databases that lead to long time and loss of useful information.Text summarization is a solution to produce an of information. Summarization has been studied for many years as a word processing tool. The purpose of a summary is to produce a text shorter than the original text so that the summary contains important information and no redundancy. So far, different algorithms have been proposed using statistical concepts and machine learning and other methods. In the last few years, deep learning has attracted a great deal of attention in many areas, including the processing of natural languages, and has achieved great success in these fields. In this research, a method based on Conditional Generative Adversarial Networks (CGAN) is presented for extractive summarization which has a Generator and a Discriminator that compete with each other in a process. The purpose of the generator is to select important sentences from the text for summarization and the discriminator's goal is to identify the generator's efficiency. In this research, after identifying the important characteristics of the text sentences, we provide it as an input to the generative adversarial network and finally, after training, the important sentences are determined by the generator. We evaluated our proposed model on two types of datasets, CNN/DailyMail and Medicine. Summaries produced in this study performed 5\\% better than competing algorithms. Text Summarization, Redundancy, Deep Learning, Generative Adversarial Networks, Extractive Summarization