With the development of various derivative products based on stock index funds and volatility, banks and investment trust companies have been focusing considerable attention on risk management. Accurate estimation and forecasting of volatility is a core issue in risk management for the evaluation and hedging of various types of funds and for establishing investment strategies. Traditional financial time series analysis techniques have been used as the primary forecasting methods for volatility prediction. This paper uses the KOSPI 200 index to compare forecasting methodologies from prior research and proposes an integrated model combining financial time series models and artificial neural networks. In terms of directional forecasting of volatility, the GARCH model from financial time series outperformed the artificial neural network model, whereas the artificial neural network model demonstrated higher forecasting accuracy than the GARCH model in terms of volatility prediction precision. Therefore, this paper suggests the possibility of simultaneously pursuing both directional accuracy and prediction precision of volatility through the integration of artificial neural network models with various financial time series models (EGARCH model, GARCH model, and EWMA model).