Directional technical indicators are often considered a useful tool for forecasting stock market trends. In this study two feed-forward neural networks with backpropagation learning are trained to forecast the directional technical indicators +DI, −DI and ADX. By combining technical analysis and fuzzy logic, a new hybrid trading system has been developed with the goal of generating early buy∕sell signals, and hence better returns. The historical price data of the S&P 500 index from the year 2001 to the year 2005 are used in this study. The results obtained by the hybrid trading system are compared against those obtained using a buy-and-hold strategy, a technical indicator “rule-of-thumb” trading strategy, and a trading system based only on the neural network model without the aid of fuzzy logic.