Predicting stock prices is an important objective in the financial world. This paper presents a novel forecasting model for stock markets on the basis of the wrapper ANFIS (Adaptive Neural Fuzzy Inference System)-ICA (Imperialist Competitive Algorithm) and technical analysis of Japanese Candlestick. Two approaches of Raw-based and Signal-based are devised to extract the model's input variables with 15 and 24 features, respectively. The correct predictions percentages for periods of 1–6 days with the total number of buy and sell signals are considered as output variables. In proposed model, the ANFIS prediction results are used as a cost function of wrapper model and ICA is used to select the most appropriate features. This novel combination of feature selection not only takes advantage of ICA optimization swiftness, but also the ANFIS prediction accuracy. The emitted buy and sell signals of the model revealed that Signal databases approach gets better results with 87% prediction accuracy and the wrapper features selection obtains 12% improvement in predictive performance regarding to the base study. In addition, since the wrapper-based feature selection models are considerably more time-consuming, our presented wrapper ANFIS-ICA algorithm's results have superiority in time decreasing as well as increasing prediction accuracy as compared to other algorithms such as wrapper Genetic algorithm (GA).