The main objective of this research is to analyse the long-term memory of the Iranian stock market and compare it with the capital market of Islamic environments. The method of this research is quantitative because financial information is used to calculate the variables. The time frame of the research is from 2008 to 2023, and the investigation of the research hypotheses has been carried out using the real data of these years. Based on the purpose of the research, the information related to the above period and the companies that were present in the securities market during this period are included in the test of hypotheses. In this research, the LSTM artificial intelligence algorithm model is used. LSTM stands for (long short-term memory) and is a type of model or structure for ordinal data that has been developed to develop recurrent neural networks (RNN). In the LSTM model, for the size of windows 30-60-90 and 180 with a horizon of 5-10-15-20, the indices of the Iranian stock market, the Malaysian stock market and the Pakistani stock market were studied. In this model, since 3 different sizes of windows have the ability to predict, therefore, all the research hypotheses have been confirmed and the stock market indices of all three countries have long term memory.