Black and Scholes(1973)發表了有名的Black-Scholes選擇權評價模型，自此選擇權評價逆成為重要的研究課題。雖然目前Black-Scholes模型已廣為各界所使用，但其股價波動度為常數之假設卻常與實際情況不節合。本文主要使用Duan(1995)所提出之GARCH選擇權評價模型，嘗試放寬股價波動度為常數之假設，以符合現實情況，並計算臺灣加權股價指數選擇權買權在存續期間每日的風險值(Value-at-Risk)，最後以穿透率(violation rate)作為評比準則，比較各風險值計算方法之優劣。本文實證結果顯示，一般而言，GARCH選擇權評價模型較Black-Scholes選擇權評價模型更能準確估算臺灣加權股價指數選擇權之風險值。Black-Scholes選擇權評價模型更能準確估算臺灣加權股價指數選擇權之風險值。

　　　　　Black and Scholes (1973) developed the famous Black-Scholes option pricing model to price the options related derivatives. The assumption of constant volatility in Black-Scholes model has been show inconsistent with the market behavior in most empirical studies. In this paper, we release the constant volatility assumption by using the GARCH pricing model developed by Duan (1995). The purpose of this paper is to evaluate the option VaR estimation performances of various Black-Scholes and GARCH pricing models for options traded in Taiwan. In general, our empirical findings indicate the GARCH models perform better than the Black-Scholes models.