Natural gas storage valuation is a challenging topic. It requires both appropriate price models and optimization models. We assume that each forward price follows a geometric Brownian motion process. Spot price also follows a geometric Brownian motion process, but its expectation changes from month to month. Since optimization models are independent of price simulations, alternative price models can be implemented in the valuation. Various models are used for storage valuation, including "intrinsic rolling with spot and forward", Monte Carlo simulation with ordinary least square regression, and Monte Carlo simulation with stochastic dual dynamic programming. The first methodology takes both forward and spot prices into account, while the other two methodologies only use spot price. The results show that many factors can impact the value of a storage, including the term structure of the forward prices, volatilities of forward and spot prices, and the operational flexibility of the storage.