Forecasting methods which incorporate appropriately chosen exogenous variables (EVs) produce enhanced forecasting performances than single variable time series methods. However, suitable exogenous variables are hardly available in practice. This study introduces a new forecasting approach, known as Wavelet Autoregressive Integrated Moving Average with WCs as EVs and Generalized Autoregressive Conditional Heteroskedasticity integrated with Wavelet Artificial Neural Network (WARIMAX-GARCH-WANN) method, to capture the data dynamics and enhance predictive power and accuracy, and, at the same time, address the challenge of non-availability of EVs. The WARIMAX-GARCH-WANN method uses Wavelet Components (WCs) extracted from the wavelet transformation of the underlying time series. These WCs are taken as conventional EVs by WARIMAX-GARCH-WANN method. Like GARCH and ARIMA-GARCH methods, the WARIMAX – GARCH-WANN method is used for high frequency time series which display nonlinear characteristics like non-constant conditional variance that hinges on lagged values of the time series. Moreover, it models frequency structure present in the data series to help achieve better performance in terms of prediction. The application of the WARIMAX-GARCH-WANN method to Wilshire 5000 Price Index commendably outperforms the WARIMAX-GARCH, WANN in terms of performance for both insample and out-of-sample forecast results. Supervisor:-Dr Ahsan ul Haq