The academic fields of computational statistics and financial mathematics have used stock index predictions as a benchmark for several decades. For the purpose of making reliable forecasts of stock prices, several distinct machine learning algorithms have been created. In this research, we use a hybrid model that incorporates both traditional CRO techniques and an ANN to predict the BS E's indices. To get a good start on our CROs, we employed the UP method to establish a large pool. The preprocessed data used for training and testing includes the daily closing prices on the BSE. The model's prediction accuracy is measured against that of a multilayer perceptron (MLP) model using the Average Percentage of Errors (APE) metric. Following this rationale, it appears that the ANN-CRO model could be a useful tool for predicting market indices.