The stock markets is always suffering from ups and downs. Investors apply their money, mostly in companies stocks, and expect a capital appreciation over time. A good investor has theoretical and practical knowledge about the investment market and can make risk assessments of a stock, analyzing previous trends and behaviors. However, due to the high number of variables that can impact the performance of a business, it is not easy to identify patterns that indicate the best time to buy or sell shares. In order to operate in this complex environment, investors have used advanced forecasting and graphical analysis tools to assist in the decision-making process. This work proposes methods to help that decision-making process in financial applications and it is based on two different strategies for buying and selling stocks. The first uses only the LSTM neural network to make buying and selling decisions. The second, in addition to the neural network, also use candle patterns to influence the decisions. In that case, the patterns will only influence when the LSTM indicates that a buy or sell transaction should take place. The strategy that uses only LSTM has proven to be effective in stocks that already showed a growth trend. When it was combined with the Candle patterns, the result was a safer strategy. It showed a lower yield, but reduced losses and the number of operations.