Accurate forecasting of stock market indices is imperative for investors, financial experts, and professionals, enabling them to make well-informed decisions and proficiently oversee their investments. This research conducts a comparative analysis of three forecasting models: RW (Random Walk), ARIMA (Autoregressive Integrated Moving Average), and ANN (Artificial Neural Network), applied to India’s prominent stock market benchmark, the Nifty fifty index. The dataset comprises daily adjusted closing prices of the Nifty fifty index spanning from January 2018 to June 2022, totalling 1106 trading days. This study employs two widely recognised error metrics, Mean Absolute Error (MAE) and Root Mean Square Error (RMSE), to evaluate the forecasting accuracy of these models. The results consistently demonstrate the superior performance of the ANN model over RW and ARIMA models, offering valuable insights for various stakeholders. The findings of this study have significant implications for academia, traders, investors, fund managers, and regulators.