The pricing of the first bond with attached warrant in the domestic capital market of China was studied. The forecasting ability of Black-Scholes warrant pricing model were examined empirically. Volatility forecasting is very important to derivative pricing, hedging, and risk management. Warrant pricing biases related to warrant strike price, time to maturity, volatility, and interest rate differential will be considered. Three different methods were employed to compare the accuracy of forecasting ability, which are implied volatility, historic volatility, and GARCH model. It is found that both the implied volatility method and GARCH model will beat historic volatility and significantly reduce the model mispricing.