Abstract: This study proposes Ant Colony Optimization (ACO) algorithms approach to determine the connection weights for Artificial Neural Networks (ANNs) and predict the stock price index. Earlier research proposed many varied models of ANN for the method of training the network, feature subset selection and topology optimization. In most of these studies, the optimum weights are not used to improve the learning algorithm. In this study, ACO algorithm is employed not only to improve the learning algorithm, but also to reduce the complexity in feature space. The ACO algorithm optimizes the connection weights between layers in neural network. This method decreases the limitations of the gradient descent algorithm. Experimental results show that ACO algorithm approach to the optimum model in compare to the other conventional models.