The stock market is an economic thermometer in every country, and news reports usually mention the financial index regarding the stock market. Therefore, predicting this variable provides bright insight into the economic situation and obtaining appropriate investment strategies. Research on stock market predictability has a long history in financial economics. While there are different opinions about market efficiency, extensive empirical studies show that financial markets are somewhat predictable. Therefore, among the methods of predicting stock returns, statistical or econometric methods based on analysing past market movements are more accepted than other methods. Artificial intelligence has also been prevalent with the increasing computing power of computers. The current research uses artificial intelligence models to predict the total index of the Tehran Stock Exchange and the index of the top 50 companies. Moreover, for this purpose, the information related to these indices has been extracted in one-day intervals for ten years. We will analyse this information with the deep learning model, one of the new artificial intelligence models.