The first objective of this study is to predict the Korean stock market volatility with the time-series model, implied volatility derived from Black-Scholes option pricing model and artificial neural network model. The second is to compare the forecasting performance among the models by the error statistics such as ME, RMSE and MAE. To examine this issue, this paper employs daily KOSPI 200 stock index data for 10 years. The monthly variance, calculated from the daily return, is used in forecasting and the benchmark of the ex-ante predictive value. Empirical results show that the artificial neural network model provides superior forecasts of volatility based on the error statistics.