The stock market is one of the most vital components of a free-market economy, as it provides companies with access to capital in exchange for providing investors partial ownership. By trading in stocks, investors have an opportunity to earn capital gain and investment income. We focus on improving capital gains by tackling the challenge of stock price prediction. In this research, technical indicators were used to predict stock price. We proposed a Tailored Ensemble Approach (TEA) to improve accuracy. In this approach, various regression models (e.g. SVR, DT, etc.) were considered as base models. Also, multiple feature sets were developed to use to train each regression model. Our ensemble approach finds the combinations of regression models and feature sets, through a validation process, that best predicts future price for each stock. Once the set of combinations are constructed for each stock, a mean ensemble system is used to predict future price. S&P 100 stocks dataset, spanning from the beginning of 2007 to the end of 2016, was used to evaluate this research. After evaluation, using various configurations, the proposed system consistently outperformed all base regression models used in this study. It was also demonstrated that a customized approach which tailors a set of regression models and feature sets based on stock and prediction period has significant impact on improving the accuracy of the stock price prediction system.