Abstarct: Nowadays, due to the expansion of societies and the increasing financial needs of human society, investing in financial markets has special importance in societies. In addition to local stock markets, Forex, as an international financial market, includes a large volume of financial market transactions. As investing is always risky, safe investing requires knowledge and a lot of time; Therefore, a trading assistant with the ability to identify suitable situations to enter and exit trades can play an important role in a safe investment. In the present study, a model for predicting the future of the financial market is offered using machine learning methods such as deep neural networks LSTM, CNN, CNN-LSTM, and random forest algorithms, support vector machine, and genetic evolution algorithm with the help of technical indicators in technical analysis of financial markets. It is offered in the form of providing buy or sell signals to the investor. The data tested in this study is a collection of 10-years data from the British pound and the US dollar (GBP / USD) in the Forex market from 2011 to 2021. One of the main features of this research is to pay attention to different investment time horizons, ie short-term, medium-term and long-term horizons by examining market conditions in the short-term to long-term timefrxames. For this purpose, we extracted a feature for each timefrxame baxsed on the Ichimoku indicator and considered these features along with other features of the dataset as the input of machine learning models. During the present study, the output of the LSTM deep learning network-baxsed model, which is a neural network for learning time series data, on a 1-day time frxame with 78% accuracy, was better than other models. Also, compared to the 1-hour and 15-minute timefrxames, the prediction for the 1-day timefrxame was more accurate.