Predicting share prices of stock market is a common topic of research in all over the world because of interest of the mass investors. However, it is a challenging issue because price is affected by external factors and has higher volatility. In this study, Artificial Neural Network (ANN) is a latest tool which has used to predict the Dhaka Stock Exchange Broad Index (DSEX) for the duration from August, 2013 to December, 2018. DSEX represents around 97% of the total equity Market Capitalization. In this paper Three Layer Feed Forward Neural Network using Back propagation learning algorithm is used. The accuracy of the model is examined by calculating Root Mean Square Error (RMSE) values. The fitted ANN model showed better prediction pattern with smaller error values.